Last Tuesday, Science and Technology Secretary Liz Kendall mystified and horrified my Twitter timeline in equal measure when she tweeted a BBC One segment about Barnsley’s status as Britain’s first ‘Tech Town’. Announced back in February, Barnsley is receiving £500,000 in funding over the next eighteen months to support AI skills training and the rollout of AI in healthcare and education.

The announcement, which amounts to roughly £2 per resident, claimed that the new “status will position Barnsley as the UK’s trailblazer”. In many ways, this epitomises technology policy in Britain at the moment: unambitious, scattergun, and communicated in rhetoric totally mismatched to the substance.

Despite the technology policy ecosystem producing extraordinary volumes of output, both the proposals and debates I see feel increasingly stagnant. No matter the ideological predisposition of the author, most reports seem to converge on similar sets of recommendations around compute, skills, sandboxes, prioritising specific sectors for support, data access, procurement reform, immigration reform, and copyright reform.

These proposals are a blend of the sensible, the unlikely, and unambitious filler. It’s often hard to escape the feeling that these reports are the product of supply-side dynamics in the think tank world rather than any discernible objective. Many of the issues holding back the tech sector now sit in other areas, such as energy or planning, but it’s also hard to write policy when the government itself doesn’t have a discernible agenda.

Back in the glory days of February 2025, Keir Starmer promised to “mainline AI into the veins” of the UK economy. The government seemed set to pass radical pro-innovation copyright reform. Investors and CEOs were queuing up to engage with ministers and announce investments. But that optimism hasn’t survived contact with reality. Copyright reform was shelved in the face of vested interest stakeholder pushback, investment announcements have dissolved into thin air, while the government’s wider agenda seems designed to deter adoption.

Much of this can be explained by how the government never really had a theory of technology-enabled economic growth. It has cycled through several different semi-theories, but each of them has run into a structural constraint of some kind that it wasn’t prepared to address. Each of these theories, however, tells us something about the policymaking process and the constraints any future government might face.

Subscribe now

The AI doctor won’t see you now

Early Starmerism, inspired by the Tony Blair Institute’s work on ‘reimagining the state’, engaged in much enthusiastic discussion of AI-enabled public services. The snappily-named AI Opportunities Action Plan, published in January 2025, stated that “AI could be the government’s single biggest lever to deliver its five missions”. There was talk of AI saving the hundreds of thousands of hours of time in the health service lost to missed appointments and huge efficiencies in areas like benefits processing. In March of that year, then DSIT Secretary Peter Kyle said that it was “almost certain” that AI would make it possible to reduce civil service headcount.

Better public services are a good thing and you can construct a second order argument that explains how they contribute to growth. For example, healthier workers are more productive and faster planning decisions unlock investment. But modest, diffuse efficiency gains are not the same thing as macroeconomic strategy. When a private firm automates a process, the productivity gain shows up directly in output per worker and in margins. When HMRC automates an element of tax processing, the gain dissipates across millions of taxpayer-hours saved, none of which are measured, and the internal cost saving may not materialise if the staff are just reallocated.

And therein lies the problem. Real efficiencies are contingent on complementary policy changes that the government isn’t in all likelihood prepared to make. When the Labour Party is polling at 16%, it’s unlikely to pick a political fight with its own base by meaningfully reducing public sector headcount. In the same way, you can’t use AI to bring down the welfare bill when the primary problem is the eligibility criteria rather than fraud. The government’s forays into reforming these criteria haven’t been promising so far.

In the end, the great efficiency agenda amounted to a handful of relatively unambitious pilot programmes (some of which have already been discontinued), and by the autumn of 2025, the government’s rhetorical energy had largely shifted elsewhere.

Pig iron production up 200%, comrades

The government has enthusiastically paraded commitments from American tech companies to build AI infrastructure in the UK, culminating in a £31 billion “Tech Prosperity Deal” announced during Trump’s state visit in September 2025. Microsoft committed £22 billion over four years, NVIDIA pledged 120,000 advanced GPUs, and OpenAI announced a UK version of its Stargate project. Starmer declared the deal would lay “the foundations for a future where together we are world leaders in the technology of tomorrow.” (On a personal note, I find this Pravda-esque style of government, which involves wielding out contextless big numbers to signal great productivity gains quite tedious.)

Six months later, the picture isn’t looking promising. OpenAI announced that it was pausing Stargate UK, citing the cost of energy and the regulatory environment. Government ministers are, of course, right when they point out that British energy costs were known to OpenAI at the time they made the commitment. Meanwhile, a number of the headline commitments from other tech companies look less impressive up close. For example, CoreWeave’s £1 billion “investment”, announced with promises of “two new data centres”, turned out to involve renting space in facilities built in 2002 and 2015. Meanwhile, the government has acknowledged that a £1.9 billion commitment from NScale was “not a formal contract, rather an intention to commit capital”.

DSIT themselves have admitted that they simply publish the investment numbers companies give them and that they make no effort to check them.

Even were all these announcements to be real, data centre hosting is probably the lowest-margin part of the AI value chain. Britain would capture employment, operational roles, and business rates on the physical facilities, but the recurring revenues from every API call would ultimately flow back to the US.

A host also needs to be able to guarantee grid connections and the ability to build – something it has yet to manage. The first AI Growth Zone, at Culham in Oxfordshire, was announced in February 2025. It has yet to begin building work and is still considering delivery partner proposals.

Even before OpenAI bailed, the government had been backing away from the data centre play, as part of its retconned evolving theory of tech sovereignty. Back in January, Liz Kendall said in a speech that “we shouldn’t try to outrun the US or China on building the most or the biggest data centres”, and should instead be focused on adoption.

Something something towns

A pure focus on adoption would have a lot to recommend it, but it runs into a combination of factional Labour politics and the wider incoherence of the government’s economic agenda.

AI-related ministerial roles have been consistently held by figures from the right of the government, such as Peter Kyle and Liz Kendall. The faction, identified with former Starmer chief of staff Morgan McSweeney, saw Labour’s path to power running through towns and suburbs, as opposed to cities.

If you’re serious about adoption driving growth, you’d target your interventions at the firms and sectors where AI adoption would generate the largest productivity gains. That means, large firms (which have the data infrastructure, management capability, and scale to deploy AI effectively) and knowledge-intensive sectors (like professional services, finance, IT, and pharmaceuticals, where AI tools are most mature). Focusing narrowly on towns is likely to miss both.

Instead of working out how to ensure mid-sized professional services firms in Leeds are integrating AI into their services, we’ve seen a series of low-value labour market interventions, framed in the language of “working people,” “local communities,” and places that have been “short-changed for too long”, presumably like Barnsley, Town of Tech.

These include £5 million ‘community funds’ around the AI Growth Zones (largely in former industrial areas) and a new Future of Work Unit (to “protect communities from the mistakes of past industrial change”), and regional “AI Adoption Hubs” to do … presumably something. I’ve also heard stories from people working in government about how they are constantly receiving requests from ministers and officials for examples of how AI will benefit average wage-earners in towns.

One step forward, three steps back

Now I’m all for the average wage-earner in towns doing better, but AI will likely only bring this about indirectly. If Britain manages to capture the economic rents from deploying AI, then we’ll be richer. We’ll be able to pay for more public goods that people in towns benefit from, while rising aggregate demand will benefit people who produce traded goods and services.

Unfortunately, the government’s own research on AI adoption found that 80% of UK businesses have no plans to adopt AI at all. This is not because until now, these businesses did not have a government hub to teach them how to use ChatGPT. It is more likely because the government’s wider policy agenda seems designed to disincentivise it.

For example, the government has hiked the minimum wage and employer national insurance contributions. The Low Pay Commission, whose evidence informs government policy on this issue, has found that employers often cover increases in the minimum wage by cutting investment.

As well as reducing firms’ financial buffers, the government is increasing the legal risk they face. If AI adoption means some roles change and others disappear, firms need flexibility to redeploy and, in some cases, make redundancies. The Employment Rights Bill, with measures like day-one unfair dismissal rights, risks making it meaningfully harder to restructure a workforce around AI.

Meanwhile, in March, the government declared that it no longer has a policy position on AI and copyright, having backed away from an opt-out model. A report following a government consultation concluded that there was “no consensus” on the issue – which was presumably the point of running the exercise. As well as creating yet more investment-deterring uncertainty, this sent a clear signal that the government will retreat from pro-AI positions when a politically vocal constituency objects.

We can also throw in the government’s online safety agenda, which is among the most burdensome and authoritarian in the western world. To signal its openness to investment, Number 10 summons in the bosses of tech companies so the prime minister can performatively berate them on camera.

Depressingly, I suspect that the majority of the people sat around the Cabinet table simply haven’t clocked that there is a tension between AI adoption and the government’s wider agenda (or lack thereof in some domains). More worryingly, I suspect a big chunk of policymakers simply don’t care. It was reported that at a cabinet away-day last July, both Ed Miliband (the Secretary of State for Energy Security and Net Zero) and Lisa Nandy (Secretary of State for Culture, Media, and Sport) reportedly grumbled about there being a session on AI. Meanwhile, Chancellor Rachel Reeves and Liz Kendall have both talked about how they don’t personally use AI, although this could be a product of an eccentric Freedom of Information ruling.

Viewed through another lens, this incuriosity is rational. If Number 10 and DSIT can’t stick to a narrative on the importance of AI for more than a few weeks or months at a time, there’s little incentive for ministers to devote that much intellectual time or attention to it. This is part of the reason that the act of cycling between theories is so damaging. If the boss doesn’t seem to believe in any one of them all that much, why should you invest any political capital yourself?

Groundhog day

A coherent theory of how technology drives growth needs to specify a mechanism: technology does X, which causes Y, which produces Z (growth). At various points, the government has cycled through a spending efficiency argument (public services), an input argument (attracting investment), and an output argument (towns). At no point do these ever join up into anything more substantial or coherent.

The previous government, for all its faults, had something approaching a theory. They believed that R&D funding would lead to breakthrough technologies, which would then be commercialised, creating new high-tech businesses that then contributed to growth.

It wasn’t supremely sophisticated, but you could at least chart a through-line through policy decisions. The Catapult network was established in 2011, modelled on Germany’s Fraunhofer institutes, to bridge the gap between university research and commercial application. Innovate UK, which funds promising technology companies, was folded into UK Research and Innovation in 2018 to align it more closely with the research councils. Meanwhile, EIS and SEIS, the tax-advantaged venture capital schemes offering 30% and 50% income tax relief respectively on early-stage investments, were expanded repeatedly.

In 2023, it published the Science and Technology Framework, which outlined priority technologies, accompanying investments, and contained specific actions to be taken by departments. The Office for Science and Technology Strategy was a serious attempt to build a vehicle for implementing this. The current government issued a revised Science and Technology Framework, which removed all of the funding commitments and specific proposals, replacing them with a list of principles. The underlying logic around strategic advantage was stripped out, in favour of bromides about industrial strategy and ‘missions’.

Some parts of the old agenda made sense, other bits were profoundly misguided. In their early years, they believed that innovation policy could be outsourced to universities, which produced the Alan Turing Institute disaster. It was only in their sunset years that they moved into institution-building mode with ARIA and the AI Security Institute. The previous government also bears much of the blame for creating the multibillion pound subsidy machine for predominantly middling to poor venture funds. And the less said about EIS the better. But there was at least an appreciation of what the machinery of government did and how it could be deployed in the service of a wider agenda.

Our current government is now attempting to revive elements of this agenda. It’s achieving this partly by compelling pension funds to invest in overvalued private markets, but also with its recently unveiled £500 million Sovereign AI fund. Around half of this money has been set aside for equity stakes, and the rest for grants. SovAI also offers other support, such as stakes, fast-track visas, and access to procurement.

From what I know about their first wave of companies they’ve backed, I’d be surprised if many of them were struggling to raise money from private investors. This leads me to suspect that their primary motivation for taking government money are the other value-adds. If I were a betting man, I imagine they’re most interested in the procurement.

This is probably the hardest part of Sovereign AI’s work, largely because the organisation itself has limited control over it. You can have the best investment team in the world (and Sovereign AI does have some great people), but they can’t compel government departments to buy products or services at scale or to make advance market commitments. They can’t instruct departments to ignore the legal advice about fair procurement that produces lengthy, incumbent-friendly competition processes. As with every other ‘theory’ here, success is contingent on other policy changes.

Other legacies of the previous government continue to perform well. The AI Security Institute’s evaluation of Claude Mythos Preview was a genuinely impressive example of state capacity. It’s unfortunate and telling that the state’s response was to write a cringe-inducing letter to businesses, telling them to set better passwords, along with other pre-AI basic cybersecurity advice. The British state created a global public good in the form of AISI, but seems more interested in appropriating AISI’s prestige to look serious, as opposed to using its insight to shape policy.

Closing thoughts

One of the reasons we seem to circle back to the same sets of solutions is that the government seems only capable of viewing innovation through two lenses: a magic force that creates jobs or moonshot ideas that are catapulted to commercial success by venture capture. The latter view is a product of over-reliance on contemporary Silicon Valley as a reference point and of spending too much time talking to universities and VCs, who are all too keen to reinforce that perception.

Anyone looking to develop a theory of technology-enabled growth should probably widen their perspective. For example, much of the most exciting work we see today is being conducted in companies that became pretty good at doing less experimental work well. Many companies built strong positions in rather dull markets, before using the profits from that work to cross-subsidize innovation. Samsung went from textiles to insurance to consumer electronics, and then to chips, while Huawei’s first product was phone switches. Microsoft is building its AI position on decades of boring enterprise software profits, while Waymo and DeepMind are the beneficiaries of a Google warchest, built on the sale of ads.

A theory of change that looked seriously at Britain’s disproportionately large share of small businesses, low levels of business investment, and sluggish technology adoption would likely yield more than asking what AI will do for hardworking families in towns.

Disclaimer: these are my views and my views alone. They are not the views of my employer, the people of trailblazing tech town Barnsley, any regional AI adoption hubs, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Subscribe now

Introduction

I’ve written before about how the UK is running contradictory technology and climate policies and how I think nuclear power is the only way out of this. If you want to understand why wind and solar are not the way forward for us, Dieter Helm has written an excellent explanation.

Early nuclear buildouts, such as the UK’s in the 1960s and France’s in the 1970s–80s, saw reactors approved and constructed within five to six years. These designs were smaller than modern reactors and significantly less complex, but had a strong safety record. Calder Hall, switched on in 1956, was built in three years and safely generated electricity until decommissioning in 2003.

Unfortunately for enthusiasts, the UK’s attempt at a nuclear renaissance has been pretty shambolic so far. Hinkley Point C, which is currently under construction, is set to be the most expensive power station built anywhere in the world ever. The project is running 16 years behind schedule and about £17 billion over budget. Sizewell C, despite being built with an identical design, is set to be even more expensive to construct, though since it has a cost-plus contract, and is 45 percent government owned and funded, it may spend less on interest payments than Hinkley, where the private sector took on all the risk and debt.

While reactors are still built at pace in South Korea, China, and the UAE, things have become slower and costlier nearly everywhere else. But the UK has the dubious honour of being the most expensive place to build a nuclear power station in the world. And it’s not even close. Much of this in my view is driven by politics rather than science.1

Against this backdrop, the government commissioned an independent taskforce earlier this year to produce recommendations on how to streamline the process. The taskforce, chaired by regulatory expert John Fingleton, published its findings today. In a rare positive Chalmermagne post, I think this is a rare example of a government review that has the potential to be transformative.

What went wrong?

To understand why this report is important, you need to understand just how bad the situation is at the moment.

There is no single regulator or unified process for approving a nuclear project. Instead, developers must secure a nuclear site licence (itself made up of several layers of independent approvals from different bodies), multiple environmental permits, and full planning consent. Each regime evolved in isolation through decades of legislation and case law. Each of the three is individually conservative, but they also overlap, resulting in significant duplication of effort and documentation.

Hinkley Point C is instructive. EDF sought to build the European Pressurised Reactor in the UK, a design created with Siemens in the early 2000s to exceed public expectations on safety. For example, it had four emergency cooling and power systems, twin thick concrete shells (one of which could withstand the impact of a fully loaded passenger jet), and a steel-lined core catcher. EDF modelled the probability of core damage at one in ten million reactor years – an order of magnitude lower than the industry standard. As a result of the UK’s regulatory process, they had to make 7,000 design changes, which added 35 percent more steel and 25 percent more concrete.

After satisfying the Office for Nuclear Regulation (ONR), EDF produced a 31,401 page environmental statement and seven Habitats Regulation Assessments. They had to mitigate any potential environmental harm that they could not incontrovertibly rule out.

The most notorious example is the hundreds of millions of pounds that EDF had to spend on fish protection, in case fish entered the intake vents for the plant’s cooling system. This included a system to retrieve and return fish and an acoustic fish deterrent that would blare out low frequency pulses to scare the fish away. Despite EDF appealing the latter on the grounds that there was no danger to any protected species of fish, a government inspector demanded that EDF press ahead on the grounds “it cannot be concluded that there would not be adverse effects”. In other words, the developer had failed to prove a negative.

Layered on top of this was a sea of other permits. For example, permission to place a diesel generator onsite during construction or licenses for dredging and soil disposal. Each of these involved separate consultations and sets of conditions. The Environment Agency’s 2023 amendment of the water discharge permit alone drew 243 consultation responses and a 120-page decision document.

Much of this documentation is the product of legal risk, rather than the rules themselves. Environmental impact assessments for major projects used to be much shorter: the 1993 Jubilee Line Extension required only a few hundred pages. Since 2001, however, the UK has been bound by the Aarhus Convention, an international treaty that is meant to secure public access to justice on environmental issues. This caps claimant costs in environmental judicial reviews at £5,000 for individuals and £10,000 for groups.

When litigation is very cheap, it encourages speculative challenges from anti-development groups that hope to catch projects out on technicalities. Developers respond by producing enormous volumes of documentation to minimise legal exposure. Hinkley has defeated three such challenges; Sizewell C has beaten two and is fighting a third.

How we fix this

Given that no one ever set out to design this system, John Fingleton and his team have done us all an immense favour by taking the time to piece together, what they call, the ‘system failure’.

They have also proposed some genuinely radical measures. The proposals in this report won’t get us down to Chinese or South Korean costs and timelines, but they’ll bring us in line with countries like France or Finland. It’s not the promised land, but it could mean nuclear costs dropping by as much as 50 percent.

There’s a lot in this report and it’s pretty accessible, even if you don’t follow nuclear. I’m going to focus on a few measures that have the potential to be particularly impactful.

Streamlining regulation and improving incentives

To stop several bodies regulating the same issue, the Review proposes designating a lead regulator for every project, which coordinates all regulatory input and resolves tensions between agencies. It also recommends simplifying responsibilities so that each category of safety or environmental issue is handled by a single authority rather than being shared.

Longer-term, it proposes creating a Commission for Nuclear Regulation. This would sit above ONR, the environmental regulators and planning bodies and act as the final arbiter on big cross-cutting decisions, rather than leaving projects to negotiate between half a dozen separate authorities.

There are also some hard-edged proposals around ensuring proportional decision-making. When regulators aren’t obliged to account for time and cost, it incentivises them to pile endless requirements onto other people. As a result, the report recommends amending secondary legislation so regulators have to consider “the cost, time and difficulty involved” when requesting changes during the licensing process. On the environmental side, developers would publish high-level cost estimates for assessments and any mitigations over £500k. The decision-maker would then have to decide whether that mitigation is proportionate and to explain why it does not discourage future development.

Subscribe now

Ending disproportionate conservatism

The Review argues that safety regulation has drifted into a pattern where the regulator drives already minuscule risks even lower, well beyond what international norms or scientific evidence require. It notes that ONR’s own dose targets sit far below natural background radiation, sometimes by two orders of magnitude, and below the internationally recommended levels used for medical and industrial radiation. This results in designers adding more systems and complexity simply to shave hypothetical doses that are already negligible.

The report cites an example I’ve dug into before about how the ONR forced GE-Hitachi to add a pointless set of extra filters into their proposed design’s ventilation system, which cut radiation emissions by one ten-thousandth the legal dose limit.

To reverse this, the Review proposes that the government – not the regulator – should define a clear national threshold for what counts as a tolerable level of nuclear risk. Once that threshold is set, ONR and the Environment Agency would be formally instructed that risks below that line are already acceptable, and that further reductions should only be sought when there is a concrete, proportionate reason.

No more praying to the environmental rain gods

The Review contains a lot of sensible recommendations on bringing sanity to the environmental assessment process. If implemented, these would cut down on the number of reviews and strip out the need to mitigate completely hypothetical risks (e.g. “you can’t prove that X species of bird wouldn’t be disturbed in Y highly implausible way”).

More radically, it offers an alternative route to the existing system. At the moment, developers normally need to mitigate any environmental effects on site. This can often produce convoluted and ineffective interventions. As the report points out:

Instead, the report proposes the creation of a new national nature recovery fund, run by Natural England. In exchange for paying into the fund, developers could set aside surveying and permitting requirements. The idea is that this fund could support genuinely impactful habitat improvements at scale, rather than focusing on site-by-site measures.

Curbing legal risk

My personal preference would be for the UK to simply leave the Aarhus Convention. Despite our exceptionally generous system of cost caps, the UN consistently claims we haven’t gone far enough. The question of membership, however, sat outside the scope of the review. Instead, the authors have done their best to cut back legal risk within this constraint.

Firstly, the Review suggests linking the cost cap to inflation. The £5,000 and £10,000 figures were set in 2013. If we’d increased them in line with inflation, they would sit at £7,000 and £14,000. Still too generous, but better than nothing.

More importantly, it recommends allowing courts to lift the cap altogether if they believe that judicial review has been abused. If you are an aficionado of planning judgements, you will occasionally see judges’ frustration shining through. For example a Court of Appeal ruling from earlier this year on a challenge to a power station described the claimant’s approach as “a classic example of the misuse of judicial review in order to continue a campaign against a development [...] once a party has lost the argument on the planning merits.” They were unable to punish this at the time, but this may change for future cases.

In October last year, a separate review into legal challenges against nationally significant infrastructure projects recommended moving to a ‘single bite of the cherry’. This means that claimants can only litigate a specific issue once. If you challenge the awarding of a nuclear site licence and are unsuccessful, you shouldn’t be able to challenge the decision to grant planning permission using the same set of facts. Despite welcoming the review, the government hasn’t enacted this yet. This is an opportunity to correct this.

The review also has some interesting points on incentives. At the moment, if a project is challenged, developers often down tools. If a permit or planning permission is revoked, this could involve spending millions to unpick work. The review suggests indemnifying the developers of critical national infrastructure so that in the unlikely event that they are defeated, they won’t have to pick up the tab for work they have already started.

What’s next?

The government will respond in full to the recommendations on Wednesday as part of the Budget. It is vital that they accept all of them unambiguously – it’s the only way of slicing through the Gordian Knot.

As one friend of mine in the industry put it: “I can’t stress enough that whether these recommendations are implemented is the difference between nuclear being functionally illegal and so requiring me to become American to build on the one hand, and on the other hand our country having a genuine chance at becoming a nuclear power again.”

There are already rumours that officials are seeking to water down the more radical ideas and they will undoubtedly face concerted lobbying from environmental groups and others.

If the government says that it’s accepted ‘the majority’ or even ‘the vast majority’ of the recommendations, the red lights should start flashing on the dashboard. As in any review, there are a small handful of mission critical policy changes, alongside a greater number of more minor tweaks. It is very easy to imagine a world in which the government accepts the latter while watering down the former.

We’ve already seen the government gesture at nuclear reform in a not entirely convincing way, so this is its opportunity to demonstrate seriousness.

Disclaimer: These are my views and my views only. They are not the views of my employer, environmental NGOs, nuclear safety inspectors, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Subscribe now

Visits from foreign heads of state see the robotic UK government comms machinery at its most Pravda-esque. During President Trump’s visit, press releases from Gov.UK GPT announced “record-breaking” £150 billion investment into the UK to “catapult” growth and “deliver opportunity for working people up and down the country”, a “world-leading Tech Prosperity Deal” will usher in a “Golden Age of Innovation”, while “British families will also see even stronger energy security” as part of a “Golden Age of nuclear”. Tractor production up 500%, comrades!

Much like the Five Year Plan, the numbers look much less impressive on closer inspection.

The £150 billion figure is a great example. £100 billion of this comes from US asset management giant Blackstone. Setting aside that 10 of the £100 billion had been announced over a year ago, the “investment pledge” amounts to Blackstone stating they plan to invest this sum in UK assets over the next decade. There’s no concrete commitment or contract. All, some, or none of this money may actually materialise.

Recently, I wrote about how the government’s tech, environment, and infrastructure policies don’t align. Announcements about new data centres are all very well, until you find that you’re unable to connect them to the grid.

A lot of people in the Progress extended cinematic universe were understandably enthused by a run of nuclear-related announcements last week, including new corporate deals and plans to streamline design approvals.

Separately, the Jensen Huang Show also came into town to dazzle audiences with a rush of AI infrastructure pledges.

So, should working people up and down the country brace themselves for the Golden Age of Nuclear and the delivery of the accompanying opportunity? Spoiler alert: not yet.

The nuclear announcements

Last week’s nuclear announcements fell under two headings.

The first was a set of commercial partnerships:

• X-Energy and Centrica – will build up to 12 advanced modular reactors in Hartlepool.

• Holtec, EDF and Tritax – will develop small modular reactor–powered data centres at the former Cottam coal plant in Nottinghamshire.

• Last Energy and DP World – will build a micro modular nuclear plant at London Gateway port.

• Urenco and Radiant – will provide £4m in HALEU fuel supply to the US, with Urenco developing an Advanced Fuels Facility in the UK.

• TerraPower and KBR – will study and evaluate UK sites for deployment of Natrium advanced reactors with integrated energy storage.

Second, the UK’s Office for Nuclear Regulation (ONR) and the US Nuclear Regulatory Commission signed a refreshed memorandum of understanding (a non-binding agreement) designed to streamline regulatory licensing.

Moving at a snail’s pace

If you want to build a small modular reactor (SMR) or a micro reactor in the UK, you need to have a reactor design that works, some people prepared to pay for it, and the regulatory approvals to build and fuel it.

The UK has never licensed or built an SMR before and only a small handful exist anywhere in the world. Rolls-Royce recently won the Great British Energy – Nuclear competition to great fanfare, with the official announcement declaring it a “new era for nuclear power”. This means it is the ‘preferred bidder’ to build new SMRs in the eyes of the government, which makes accessing subsidies easier, but doesn’t actually expedite any approvals. The company currently envisages connecting its first reactors to the grid in the mid-2030s.

This seems appallingly slow. Rolls-Royce has long manufactured pressurised water reactors. While there will be important technical differences relating to scale and uranium enrichment levels, the company should have the supply chain, skills, and the safety culture to build an SMR faster than anyone else. It recently reached the third stage of the ONR’s Generic Design Assessment process, ahead of schedule. The GDA is essentially a way for developers to get buy-in from the regulator to their design before they apply for a licence for a project on a specific site.

So why is it taking another decade?

In short, the UK’s current policy is that a small reactor should be regulated in the same deeply conservative and inefficient way as a gigawatt-scale reactor. The ONR’s director of regulation for new nuclear reactors, has said that “there isn’t any difference; the regulatory regime is technology agnostic”. The notion that a 20-megawatt reactor needs to be subjected to the same scrutiny as one like the 1.6-gigawatt reactors being built at Hinkley Point C, despite the significantly smaller risk profile, seems perverse.

Unfortunately, the regulatory regime for UK nuclear reactors is among the most conservative anywhere in the world, where new nuclear power isn’t actually banned. A reactor will go through a process that will last a minimum of eight years before they can start work on construction would last 18 months for a comparable project in South Korea.

Much of this stems from the UK’s unusually conservative interpretation of international radiation safety principles. For example, the UK instructs developers to reduce risk “as low as reasonably practicable” (ALARP). This means that any measure that is not “grossly disproportionate” must be introduced. The UK refuses to have reactor, design-specific rules, or ‘safe’ level of radiation, which means that developers are often left trying to second-guess an inspector’s interpretation of rules, past precedent, international best practice, and … vibes.

The ONR has an extended track record of requesting trivial but expensive changes to designs that function in other countries, whether it’s unnecessary amendments to filtration systems or huge increases in the volume of concrete and steel.

I’ve written elsewhere about the poor evidentiary basis for measures that drive down already trivially low radiation doses.

In the absence of significant regulatory reform, things begin to look tough for some of the projects involved in this announcement.

Roughly 80% of the reactors operating in the world are pressurised water reactors. Rolls-Royce is designing a PWR, while EDF is building PWRs at Hinkley Point C and Sizewell C. Sizewell B, the last reactor to be built in the UK, is also a PWR.

PWRs basically work like old coal or oil plants, but with a different fuel: water is boiled to produce steam, which spins a generator. This is pretty straightforward, but a lot of the heat goes wasted. If you want to read more about reactor designs, take a look at this.

A number of SMR designers are attempting to build more efficient reactor designs. X-Energy, who are partnering with Centrica, are trying to build a high-temperature gas-cooled pebble-bed nuclear reactor. In one of these reactors, the fuel is packed into thousands of tennis ball-sized graphite spheres. Helium gas flows through the spaces between them to carry away heat. Helium is chemically inert and remains gaseous at high temperatures, meaning the reactor can operate safety at very high temperatures, significantly boosting efficiency.

This is technically impressive, but phenomenally challenging. The one such facility operating in the world is a demonstrator in China, where the nuclear sector is nationalised and given exceptionally favourable regulatory treatment. The only other pebble-bed reactor built with a similar design was a 1969 experimental plant in West Germany. Plagued by technical failures and extended outages, the small 15-megawatt unit was dubbed ‘the shipwreck’ in German nuclear circles, and it was decommissioned in 1988 after fewer than 20 years’ service.

I’m not saying that companies shouldn’t build high-temperature gas reactors or that the X-Energy design is bad. But in the absence of sweeping reforms, it seems implausible that this design could successfully run the gauntlet of UK nuclear regulation and be built before the 2040s.

The company has only completed one stage of pre-application engagement with the ONR. This tier is a single day event with the regulator, where they explain the process to a prospective applicant. This hasn’t stopped the company claiming that work will start in 2026 for deployment in the mid-2030s – roughly the same time frame as Rolls-Royce.

I’m going to skip over the TerraPower/KBR announcement, as it is a very vague commitment to think about building an experimental design in the UK. It feels like corporate marketing the government has bundled into an announcement to bulk it out.

The Holtec/EDF and Last Energy/DP World announcements seem like altogether more serious proposals. The Holtec design has passed stage one of the Generic Design Assessment, while Last Energy has cleared the Preliminary Design Review (the last state of engagement before making a formal site licence application). Both companies have therefore meaningfully engaged with regulators in the UK. They are also both building PWRs, which means that they are taking on significantly less technology risk than X-Energy.

In a sane world, it should be possible to build these reactors quickly, but projects still potentially face an eight year march through the UK nuclear site licensing and planning system. In essence, the government remains the biggest obstacle to its own golden age.

These points aside, the individual projects also have problems to navigate. For example, EDF and Holtec picked the Cottam site for a few different reasons. As a former coal-fired power station, it has some obvious advantages from a planning standpoint and a substation for grid connections.

Unfortunately, for EDF, the substation is taken. While they used to have a grid connection at Cottam, the capacity has now been assigned to a solar and battery project. This means that they will likely have to join the same decade-long queue as any other project.

Moreover, Cottam is not a coastal or estuary site. This means EDF and Holtec will need to rebuild the old cooling towers, which were demolished … a month ago. The alternative is directly tapping into the Trent for cooling, but this would involve an unenviable political and legal battle with environmental groups.

The Last Energy/DP World project is for a significantly smaller reactor (20 megawatts versus 300 for Holtec) and the design can use air cooling, which means it doesn’t have to navigate the tower versus river conflict. The project is being pitched as a private wire project, which means that the reactor will directly power the London Gateway port that DP World, a multinational logistics company, operates.

But that doesn’t make things easy. In a country like South Korea, or France during its rapid buildout, once a design is deemed ‘safe’, it’s much easier to build it at multiple sites as part of a ‘fleet-based’ approach. In the UK, you have to navigate the process on a project-by-project basis. The DP World project is Last Energy’s second proposed site. They still need to secure approval on their first planned project in South Wales.

Mutual recognition?

It’s against this backdrop that we should treat the ONR’s refreshed memorandum of understanding with its American counterpart with a degree of scepticism. The agreement states that the two wish to cut approval times for new reactors, aiming to finish design reviews in two years and site permits in one. To do this, they’ll divide up review work, make use of each other’s past assessments, and, where possible, accept one another’s conclusions.

The memorandum itself notes that regulators in both countries have had a cooperation agreement since 1975 and they, in fact, signed another agreement last year (along with Canada) to cooperate on advanced reactor approvals. This document already contained a number of commitments for the regulators to facilitate information-sharing and develop shared regulatory approaches that mean the approvals process in one country will produce documentation that helps answer the questions raised by a similar process in another.

The only novelty is the non-binding time targets and some slightly stronger language on mutual recognition. Even then, the ‘mutual recognition’ will only come after “appropriate due diligence”, which gives regulators essentially unlimited leeway to negate anything that the agreement might be hoping to achieve.

Real mutual recognition would be a welcome step forward, but nuclear regulation is bad … basically everywhere in the western world. Vogtle 3 and 4, the newest US commercial reactors, are among the most expensive ever built. Better than the UK’s attempts, but still dismal.

Aligning ourselves with regulation that allows nuclear power to be built at a marginally less catastrophic price is an improvement, but it’s not a substitute for rethinking regulation from first principles. If you want to read some more transformative ideas, my friends at Britain Remade recently published a serious playbook that covers science-based radiation rules, measures to streamline planning and environmental rules, and steps that would unlock a ‘fleet-based’ approach.

Spinning out of control

If there are question marks over the golden age of nuclear, the AI side of the UK’s great infrastructure build is raising some interesting questions.

The UK Government and NVIDIA jointly announced that the company would be deploying up to 60,000 GPUs in UK-based data centres by the end of 2026, with the total number potentially reaching 120,000 over the next few years. Tom Bristow at Politico dug into the numbers and found that the NVIDIA and DSIT press offices got a touch excited and the number was closer to 30,000.

During his visit, Jensen Huang said that nuclear power and gas would be required for the UK to meet its AI ambitions. In our last piece, we covered some of the uncertainties hanging over how AI Growth Zones would be powered, with some in government lobbying to limit the role of fossil fuels and instead rely on wind and solar. We now have a clearer idea, with the newly announced one gigawatt ‘Stargate UK’ project relying on diesel generators for back-up. Given the UK government shows no sign of backing down from its highly ambitious plan to transition 95% of UK energy generation to green sources by 2030, something is going to have to give.

The only hope is that the UK Government’s Nuclear Regulatory Taskforce follows through on its punchy interim report with a set of sweeping proposals for reform and the government takes them seriously.

Otherwise, it’s going to take more than Pravda headlines for the UK to back itself out of this one.

Disclaimer: These are my views and my views only. They are not the views of my employer, any West German nuclear engineers, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

"Well, in our country," said Alice, still panting a little, "you'd generally get to somewhere else – if you run very fast for a long time, as we've been doing."

"A slow sort of country!" said the Queen. "Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!"

Lewis Carroll, Through the Looking-Glass

Introduction

Back in January, Keir Starmer pledged to “mainline AI into the veins of this enterprising nation”, as he unveiled the AI Opportunities Action Plan, a report authored by entrepreneur and investor Matt Clifford, who was then serving as the Prime Minister’s AI advisor.

The snappily-named Opportunities Action Plan set out a series of sensible measures across infrastructure, sovereignty, procurement, and public sector adoption that likely tested the limits of what the Treasury was prepared to fund.

But since the announcement of the plan, the ambition of AI announcements from the UK Government has steadily slowed, while it picks up elsewhere in the world.

In the last few weeks, the US Government has published its own AI Action Plan, which among other things, contains 90 different actions spanning compute, workforce, procurement, and security. In March, the EU announced six additional sites for its high-performance computing clusters (dubbed ‘AI factories’), a month after the French had unveiled a €109 billion investment plan in French AI infrastructure. Meanwhile, the US-Gulf cooperation on compute only appears to be deepening.

At first, the UK appeared to be powering ahead on AI policy. In 2023, the country was first to establish a dedicated state-backed institute focused on evaluating the safety of frontier models. The AI Security Institute is a rare example of a new government institution that is adequately funded and able to attract top talent.

AISI has a clearly focused mission, so it can get on with its work with relatively little involvement from the rest of the government. If you’re evaluating the capabilities of a frontier model pre-deployment, the pool of stakeholders is pretty narrow (you, the people building the model) and your levers are pretty defined (asking them to make changes following evaluation).

Unfortunately, the UK has now plucked all of the low-hanging AI policy fruit. Every other important intervention requires navigating a world of departments with conflicting missions, regulators with limited interest or expertise, private actors, and legal roadblocks. This manifests itself acutely across areas like energy, infrastructure, defence, healthcare, and education.

Thanks to a combination of fragmented authority through endless outsourcing to quangos and various esoteric acts of financial engineering, the British state is particularly poorly placed to handle these. And this matters. Considering the growth in both capabilities and adoption, standing still and basking in the glory of things you set up in 2023 means falling behind.

To illustrate this, I’m going to focus on one element of the UK’s AI ambitions - its Compute Roadmap and ambition to establish ‘AI Growth Zones’. I’m not looking to debate the rights or wrongs of the specific proposals or targets here (although I’m sure AI infrastructure is a theme I will return to another time), more the extent to which the government has the ability to meet its own stated policy goals.

Subscribe now

The devil isn’t in the detail

A significant portion of the AI Opportunities Action Plan was devoted to the foundational infrastructure for AI. The two most important proposals for our purposes are instructing the government to expand the capacity of the AI Research Resource (a set of publicly-funded supercomputers available to researchers) 20x by 2030 and to establish ‘AI Growth Zones’. These zones would match up private AI infrastructure providers with land and access to clean power through streamlined planning. So far, DSIT is yet to publicly name a private partner for a growth zone, six months after the call for expressions of interest first opened.

The Department for Science, Innovation, and Technology (DSIT), which published the Action Plan, followed up with the UK Compute Roadmap. This set out the specifics of the proposal, including a goal of six gigawatts of AI-capable data centres, including one site of more than a gigawatt. It also includes a headline pledge to invest up to £2 billion in sovereign compute by 2030.

These six gigawatts presumably have to be powered by something and connected to the grid. They also translate into fifty more terawatt-hours per year – the sort of scale that requires binding power contracts and grid reinforcements.

Instead, the plan only commits to ‘exploring’ new power options like small modular reactors or onsite renewable generation, while applicants only need a ‘pathway’ to 500 megawatts of capacity. The roadmap is silent on unglamorous enablers like planning reform or grid connections.

Why is DSIT so quiet on these points? In short, because it has little power to do anything about most of them. It can spend up to £2 billion on compute itself, a relatively trivial amount, especially as £750 million of this is being allocated to restoring a single project the same department cancelled last year. But the rest of the plan amounts to a suggestion note to the rest of the government.

To demonstrate why, let’s walk through an example.

The will to power

A potential solution to the government’s missing six gigawatts could lie in bringing more nuclear onto the grid. Nuclear is clean, reliable, and could be cheap in theory. The government said earlier this year that it wanted “new nuclear technologies such as small and advanced reactors for the first time, providing flexibility to co-locate them with energy intensive industrial sites such as AI data centres”.

In June, Rolls Royce’s small modular reactor design won the competition to be selected as Great British Energy – Nuclear’s preferred bidder to build the country’s first small reactors.

Considering the government’s rhetoric about “ripping up the rules”, an outside observer would likely interpret this as meaning that Great British Energy, the government’s newly minted and publicly owned energy company, is going to buy some nuclear power stations. The government has found the reactor design it likes and is committed to streamlining the rules, so Rolls-Royce should be able to get building sharpish.

In the UK, the government does not buy nuclear power stations. It partners with a developer and creates a special ring-fenced project company, which has to win over equity backers and debt lenders. These will typically be a blend of government, pension funds, utilities companies, and large institutional investors. Raising the money can take years, investors often get cold feet, and negotiations with government over funding models have resulted in entire projects being abandoned.

Investing in a UK nuclear project is not an immediately attractive proposition. With the exception of countries that have outright banned new nuclear power, the UK has the world’s most conservative nuclear regulator, alongside stringent (and frequently duplicative) environmental consents. Developers also have to be gracious and fund up to three legal challenges against their own projects, thanks to a convention the UK should probably leave.

To derisk this proposition, the government has cycled through different models. For Hinkley Point C, currently under construction in Somerset, the government embraced the Contract for Difference. Here, the developer receives a fixed price for every megawatt-hour for 35 years, guaranteed by a government-owned counterparty. If the market price is below the guaranteed strike, consumers make up the difference through levies on bills.

For Sizewell C and future large builds, the government switched to the regulated asset base model. Under this system, nuclear developers begin receiving revenue from consumers during construction, years before the plant produces electricity. Consumers, through their bills, effectively provide an advance cashflow to lower the developer’s cost of capital. If the project overruns or faces delays, regulators adjust the “allowed revenue”. The result is that construction risk, traditionally borne by investors, is transferred to the public, while investors enjoy a regulated, inflation-linked return once the plant operates.

No successful nuclear rollout in world history has ever been based on this bizarre mix of project companies, government-owned settlement bodies, revenue contracts, and capped liability schemes. All of this financial plumbing exists to keep nuclear off the public balance sheet and to help preserve the fiction that the government is sharing the risk with the private sector, even though it is ultimately the lender and insurer of last resort.

In short, the government is responsible, but not in control.

Winning the SMR competition doesn’t guarantee Rolls-Royce a single order. Instead, it will find it easier to access money awarded to Great British Energy. When it comes to any physical reactor builds, the whole private fundraising process and hunt for investors will need to be re-run project by project. The competition itself took ten years of behind-the-scenes dithering by the government before it got off the ground.

Financing aside, the Rolls-Royce reactor design still has to make it through the gauntlet of UK nuclear licensing. This sits outside government in an independent regulator and can take four to six years, versus the 12 to 18 months you might expect in a country like South Korea. At the moment, ministers have limited power to speed up approvals.

It’s not for nothing that the interim report from the government’s taskforce reviewing UK nuclear regulation found that “the system is perceived to be unnecessarily slow, inefficient, and costly” and indicated the need for a “radical reset”.

Even if a Rolls-Royce SMR project found financing, the prospect of it being approved, constructed, and plugged in within the next decade is remote.

Missed connections

Let’s imagine all of the above was fixable. Congratulations, you’ve got regulatory approval for your power source and financing lined up for your data centre. You’re almost on your way to meeting the government’s compute and clean power targets.

Or are you?

There’s another obstacle that neither No 10 nor DSIT can help you with. This one involves Britain’s electricity system.

At the top of the system, sits the high-voltage transmission network. This moves power around the country. In England and Wales, this is owned and operated by National Grid Electricity Transmission, in Scotland, by two other companies. These firms are private utilities. Ministers do not have a guaranteed right to instruct them prioritise one project over another, any more than they can tell BT which broadband line to lay first.

These companies operate as regulated monopolities. Ofgem (the UK’s independent energy regulator) sets rules through multi-year price control frameworks currently called RIIO-2 (short for “Revenue = Incentives, Innovation, Outputs”). Every five to eight years, the network companies submit detailed business plans to Ofgem setting out what assets they want to build or upgrade, and why. Ofgem reviews these plans, consults, and then decides what investment is justified, performance targets will apply, and return on equity the company is allowed to earn.

If a project isn’t an Ofgem-approved plan, the company cannot simply build it because ministers or customers want it. If it does, it has no automatic right to charge consumers for it.

Sitting below the high-voltage transmission network are distribution networks, which carry electricity at lower voltages into towns, businesses, and homes. These are controlled by a mixture of privately-owned distribution network operators (DNOs) that act as regional monopolies, and newer independent operators (iDNOS) that build local distribution networks for specific developments.

For a moderate-sized data centre in the tens of megawatts region, the DNO is the critical gatekeeper. Even at Growth Zone scale, once bulk power has been brought from the high-voltage transmission system down into a form that can be used locally, a DNO or iDNO is still needed to build the ‘last mile’ of networks. This means that they can act as bottlenecks. Some DNOs have already had to tell developers in congested regions, like West London, that they cannot reinforce their networks quickly enough. iDNOs can sometimes ease this by building infrastructure of their own, but they still depend on connection points into the DNO or transmission system.

But when you are talking about 500 megawatts or more, the primary bottleneck is the transmission interface. Only the monopoly transmission owner can build the high-voltage substations and pylons needed to connect loads of this size. Developers must join a national connection queue run by the National Energy System Operator (NESO), a government-run body which coordinates with the transmission companies. The queue for connections is so bloated with speculative projects that it runs into the 2030s. Historically, it operated on a first-come, first-served basis. NESO is now implementing a reform programme to prioritise serious and nationally significant projects.

Even if a project moves forward, building the physical connection is slow. A new transmission-level substation and associated reinforcements typically take seven to ten years.

This is why some developers propose co-locating generation – for instance, a gas turbine or one day a small nuclear reactor – directly with their data centres. The idea is to produce power on site rather than wait for the grid. In practice, even these ‘self-sufficient’ sites still need a grid connection, for two reasons.

First, the industry standard for a data centre is 99.999% uptime. In short, it cannot afford to go dark because an on-site plant or storage system fails. It needs backup from the national system.

Second, surplus electricity from the private generator has to go somewhere when the data centre is not consuming at full tilt.

One idea under discussion is to allow Growth Zones to build private high-voltage substations and connection equipment to link their generation and demand directly to the transmission system, bypassing the need to wait for the monopoly transmission company to build one.

The one snag is that this is currently illegal. Private actors cannot directly connect to the transmission network. There appear to be no official moves afoot at the time of writing to change the law or create a carve-out under the Electricity Act to make this happen.

None of this is even accounting for the UK’s famously efficient planning system…

Fight the power

This planning point isn’t trivial, because our extra six gigawatts runs into another challenge - namely, the UK’s Net Zero ambitions. NESO, Ofgem, and the Department for Energy Security and Net Zero (DESNZ) are pursuing a policy called Clean Power 2030. The goal is that by 2030, 95% of Britain’s electricity generation should come from clean sources.

NESO forecasts electricity demand rising by about 11% in 2030, mostly from electric vehicles and heat pumps. If NESO’s assumptions held, DSIT’s six gigawatts of baseload power from the AI data centres would be responsible for 18% of all UK electricity consumption (more than twice NESO’s forecast).1

In essence, the government’s clean power and AI goals are flatly incompatible.

To attempt to square this circle, officials seem likely to decide that AI Growth Zone developers will not be able to use gas cells as an interim power source. This is good for clean power, but fatal for AI Growth Zones.

If the legal obstacles on private generation described above had been eased, then gas might have at least made the idea of an AI Growth Zone theoretically feasible, while we wait for the SMR revolution to sweep the land. The alternative is powering an AI cluster with solar panels, but given the UK’s entire installed battery storage capacity could only power a one gigawatt scale cluster for a few hours, this seems unwise given our weather.

To bring this tension into sharp relief, we are currently witnessing a face-off between DSIT and DESNZ, with a two gigawatt data centre proposal seemingly colliding with a planned hydrogen plant planned for a nearby plot of land. The data centre has received planning permission already, but the BP, the developers of the hydrogen plant are pushing Ed Miliband, the Secretary of State at DESNZ to sign the development consent order for their plant. This would include compulsorily purchasing the land. It remains unclear how these two projects will coexist, and Number 10 seems content to sit on the sidelines and let the inevitable legal battle resolve it.

The centre cannot hold

The UK has a government department with ‘science’ and ‘technology’ in its name, but it’s not clear why it exists. The UK’s future AI adoption or innovation will largely not be determined by questions by pure technology policy. Instead, it will be influenced by areas like fiscal policy, defence, education, and energy.

A strong central body capable of compelling coordinated activity across all of these would be one thing. But this is not what DSIT does. DSIT can publish strategies suggesting that the departments or regulators that oversee these areas take certain actions, but it has no power to make them do anything.

Alongside imploring British internet users not to download VPNs, DSIT’s primary activities seem to be shuffling inconsequential pots of money around and organising roundtables.

While there are many talented individuals working there, the current institutional setup is not a serious vehicle for delivering serious change.

Even where DSIT does have authority, it uses it ineffectively. DSIT theoretically has meaningful power over scientific research in the UK, but has opted to outsource this work to the UK’s research councils. These bodies, which wield an annual budget of nearly £9 billion, have presided over the UK’s declining standing in international research rankings. Angela McLean, who heads the Government Office for Science, which sits inside DSIT, described this decline as “actually a fantastic thing”, as it was proof that other countries were improving.

A cultural indifference to the UK’s relative decline is a poor foundation for a scientific renaissance.

Perhaps the most frustrating point is that none of these challenges are insoluble. An energetic prime minister, who concluded that there were a series of connected challenges across departments that were collectively serving as a bottleneck on critical capabilities, could push through them with sufficiently energetic leadership. Ministers could challenge officials who tell them action is too difficult or risky. The law could be changed to make regulators work together. Britain is, in many ways, a very centralised country. If the executive is determined to run the country, it can just do things.

But the overwhelming sense I get in my interactions with politicians and civil servants is that the country’s governing class has given up. Westminster and Whitehall have been consumed by a nihilistic passivity.

If you’re working to change anything, the consistent feedback you receive is that ministers are only interested in ‘quick wins’ and that you should avoid anything that involves primary legislation. The solution to every problem is a new taskforce or appointment. These taskforces individually may have a great deal of merit (the nuclear taskforce’s early work has been stellar), but this is not a sustainable or scalable way of governing them.

A government, one year into a five year term and with a majority of 150, that takes this approach is bound to fail.

Disclaimer: These are my views and my views only. They are not the views of my employer, regulators with conflicting mandates, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Subscribe now

Introduction

Back in March, I published a long-read exploring what went wrong with the Alan Turing Institute, the UK’s national AI and data science centre. The piece argued that a lack of clear goals, poor leadership, and the decision to outsource its operation to a coalition of universities had created an underpowered body that advised the government poorly. The only functional part of the institute was the team focused on defence and national security. This team drew less on academia and the world of research grants, in favour of partnering directly with the national security community.

After I published the piece, I received a range of responses. Many current and former members of Turing staff wrote in to express their wholehearted agreement, while a handful of academics accused me of being involved in a ‘technolibertarian’ conspiracy with Tony Blair. Jean Innes and Doug Gurr, respectively the CEO and Chair of the Turing, gave an interview that reaffirmed the institute’s present approach and didn’t address any of the criticisms. The story then went quiet.

Suddenly on 3 July, Secretary of State for Science, Innovation and Technology Peter Kyle sent a letter to Gurr, demanding widespread changes. This followed a meeting with the institute’s leadership in May. The letter praised “the existing excellent work” being done on defence and said that “defence and national security projects should form a core of the ATI’s activity” and that “the ATI’s current non-defence activity would need to be reoriented to support this renewed focus and strengthen the UK’s sovereign AI capabilities”, with leadership “that reflects the institute’s reformed focus”. I’ve added the entire text of the letter and any other correspondence referenced in this piece in an appendix at the bottom, as I realise not everyone is on X.

This is nothing short of a call for the institute’s current leadership, along with much of its current work, to be cleared out. However, as far as the Turing’s management is concerned, it’s business as usual.

Over the past week, I’ve spoken to Turing staffers, exasperated officials, and research world insiders. The picture that emerges is of a leadership determined to batten down the hatches and avoid meaningful reform. Worse still, Innes and Gurr may not even be behaving irrationally.

Please disperse, nothing to see here

When you see Peter Kyle’s letter printed in full, the irritation with the Turing is hard to miss. However, the government did not publish the text and media reporting tended to focus on the handful of quotes Politico (who were the first to get hold of the letter) opted to publish. This, unfortunately, played into the Turing leadership’s hands.

In an internal message sent to all staff the next day, Jean Innes reassured the team that the media was “quoting selectively” from a letter that was “collegiate, not critical”. Innes argued that the letter was in fact “endorsing the Turing 2.0 approach, noting our crucial role in the UK’s AI transformation, outlining the strength and impact of our existing work”.

This was an egregious misrepresentation from Innes, who was likely banking on the full text of the letter never seeing the light of day. It is true that the letter acknowledges that the institute has produced good work. It is, however, an act of olympic-level contortion to describe it as an endorsement of the current strategy.

Despite Innes and Gurr positioning it as a revamp in their FT interview, the Turing 2.0 strategy was adopted in 2023. It committed the institute to focussing on three ‘grand challenges’: health, environment and sustainability, and defence and national security. A letter calling for two of three focus areas to be scrapped can in no way be described as an endorsement of the current approach, nor is a letter saying that “further action is needed to ensure the ATI meets its full potential” conceivably “not critical”.

This would seem obvious to anyone attached to reality. But Innes seized on one sentence, in which the minister praised the institute’s work on weather prediction, as evidence that “our high-impact missions are increasingly interlinked”. The line the institute appears to be taking is that i) the government likes our defence mission, ii) all our missions are interlinked, iii) our missions are all therefore defence, iv) nothing needs to change.

The full quote from the Secretary of State’s letter was that “impactful existing work on weather prediction and fundamental AI research would align well, albeit with a renewed emphasis on, and specific linkage to, the UK’s defence, national security, and sovereign capability”. The Secretary of State is outlining a single example of work from another of the Turing’s missions that could be repurposed, not suggesting that all the institute’s current work can or should be preserved.

Based on conversations with officials and people familiar with Peter Kyle’s thinking, DSIT does not believe that the Turing’s leadership is taking the letter seriously.

Spinning out of control

Since late last year, the Turing has been running a consultation on making 143 staff redundant and has been gearing up to slash a quarter of its research projects. Throughout this entire period, the leadership team has maintained the pretence that everything is functioning as it should.

At the end of last year, 93 Turing staffers signed a letter to the Board of Trustees that criticised “the CEO’s failure to make decisions, communicate clearly, or take responsibility for the Institute’s direction” and warned that the wider leadership team’s “actions are jeopardising the viability of the Institute as a whole”. They point to how slow decision-making had led to contracts going unhonoured and potential funding falling through. The letter noted that “while ELT [the executive leadership team] continues to assure us that everything is fine, this message is increasingly at odds with the reality we observe across the Institute”.

Innes and Gurr’s response so far suggests that the leadership remains committed to its long-term strategy of denial. Multiple current Turing staffers have described leadership’s communications with staff as tantamount to “gaslighting”.

On Thursday, Gurr joined an all-team meeting and reiterated Innes’ internal message and delivered remarks that were “light on accountability” and “delusional”. While he stopped short of explicitly backing Innes, Gurr suggested that the ministerial letter should be viewed as good news and proof the institute was “exactly where we should be”. He labelled the implicit criticism in the letter as “perfectly standard questions”. In his rendering, the letter was not a criticism of the Turing, but a reflection of the government increasingly prioritising defence spending. Gurr also welcomed an apparent government commitment to allocate more money to defence and security research – something that was not in the Secretary of State’s letter and which no Turing sources believe to be true.

Gurr stressed the importance of independence and made it clear that the institute would continue with its health and environment work. As he spoke, questions piled up in the Zoom chat. Gurr is said to have taken two grudgingly, before making his excuses and exiting the meeting.

Haldane-ing onto power

To an outsider, Innes and Gurr’s response to the government may seem like a kamikaze-like provocation, but there’s a chance it might work.

The Turing is an independent body. Ministers can share their opinion on how it should be run, but have no direct ability to pick its leadership or set the strategy. Real power sits with the five founding universities and the Engineering and Physical Sciences Research Council (EPSRC), the Turing’s primary funder.

The EPSRC sits under UK Research and Innovation (UKRI), the UK’s research funding agency. The Secretary of State does have the power to issue directions to UKRI, including setting conditions on funding, but there are certain limitations. UK research policy is governed by the Haldane Principle, the notion that researchers themselves, rather than politicians, should decide how research funding is spent. The idea dates back more than a century and is currently enshrined in the 2017 Higher Education Act:

In practice, this means that once UKRI’s overall strategy has been approved by the Secretary of State, the research councils can theoretically ignore any requests made by politicians that don’t come with new money attached. Although the EPSRC is profoundly unhappy with the Turing’s current performance, and as its primary funder can take a decisive position on reform, it may be reluctant to set a precedent of rethinking funding decisions in response to political pressure.

As one insider from the research world put it: “The tactic Jean and Doug are using is to cry foul on the Haldane principle and independence, and to try to play EPSRC off against DSIT. EPSRC don't have a clue what they actually want to achieve with the ATI money, but will defend their right to manage from DSIT overreach.”

If the EPSRC doesn’t bite the bullet on this occasion, it risks creating a limbo. This would see a national AI institute operating along lines the government doesn’t want, propped up by a funding body that thinks it’s dysfunctional. The only winners in this scenario are the institute’s underperforming senior leadership team. Questions also have to be asked about the EPSRC’s own oversight, with the organisation seemingly not exercising its right to pick a member of the Turing’s board of trustees following their last representative’s retirement in the summer of 2023.

It’s all faintly awkward for the government, which appointed Doug Gurr interim chair of the Competition and Markets Authority in January 2025, after ousting Marcus Bokkerink. Any overt public clash may cast doubt on the wisdom of that appointment. Meanwhile, Patrick Vallance, now Minister for Science, sat on the panel that appointed Jean Innes CEO in 2023. At the time, Vallance described Innes as “the outstanding candidate to provide leadership of the Turing”. This may explain why, back in May, when pushed on the difficulties the institute was facing in a parliamentary question, he provided an uncritical recitation of the leadership’s script on Turing 2.0. Demis Hassabis, the co-founder and CEO of Google DeepMind, had volunteered to help the Turing on its 2023 CEO search, but, according to multiple government sources, his offer went unanswered.

Closing thoughts

The challenge of reforming the Turing illustrates the consistent shortcomings of an era in UK innovation policy. In the halcyon days of 2010-2015, under then science minister David Willetts, government assumed that agile, outsourced ventures – backed by competitive grants and co-funding from industry – would be more effective than vertically-integrated labs. It concentrated funding in a few elite universities, believing that excellence would spill over, and relied on partnerships and open data to drive commercial uptake.

While elements of this philosophy lives on in the Turing, the Catapults (government-funded innovation centres designed to accelerate research commercialisation), and individual UKRI programmes, the centre of gravity has shifted towards state-anchored, strategic capability-building. The AI Security Institute, for example, is a fully government-owned body with an explicit sovereign capability mandate. If, in a parallel universe, it had been established six or seven years earlier, it would likely have been governed by a coalition of universities and encouraged to seek paid projects with industry.

In short, if the government considers a capability important and wants it to achieve something specific, it has to pay for it directly. The alternative is a national institute over which the nation has no control.

Disclaimer: These are my views and my views only. They are not the views of my employer, the technolibertarian mafia, Tony Blair, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Peter Kye’s letter to Doug Gurr:

Dear Doug,

Following our meeting on 16 May, I am writing to outline my vision for the future of the Alan Turing Institute, with a view to ensuring it supports the government’s AI ambitions and meets the evolving needs of the nation at this critical time.

The AI Opportunities Action Plan, and the £2bn of investment now dedicated to delivering that plan, are a testament to the scale of government’s AI ambitions and our belief that AI presents an opportunity to transform the UK for the better. The national institute for data science and AI will hold a critical role in this transformation, but to fulfil that role the ATI must continue to evolve and adapt to the needs of the ecosystem.

While I recognise the progress the leadership has made towards reform, it remains clear that further action is needed to ensure the ATI meets its full potential. Whilst acknowledging the ATI is an independent organisation, its current reliance on public funding means that it is imperative that this resource is being maximised and is delivering value for money for citizens. As such, I am writing to propose a set of changes that will strengthen government’s confidence in the institute and its capacity to deliver for the country.

Boosting our AI capability is critical for growth as well as our broader national security. The Prime Minister has made clear his commitment harnessing the benefits of AI, and to further developing UK defence capabilities through a sustained increase in defence spending. There is an opportunity for the ATI to seize this moment. I believe the institute should build on its existing strengths, and reform itself further to prioritise its defence, national security and sovereign capabilities.

Expanding on this proposal, I would first underscore the vital role the ATI can play as a national security asset. I am aware of the existing excellent work undertaken by ATI in this domain through the Centre for Emerging Technology and Security, the Laboratory for AI Security Research, and the Defence and National Security grand challenge. However, there is potential to cement its place as an irreplaceable asset for the UK and our allies. Moving forward, defence and national security projects should form a core of ATI’s activities, and relationships with the UK’s security, defence, and intelligence communities should be strengthened accordingly.

With this renewed purpose, the ATI should focus on applied research, adoption and commercialisation, in line with the conclusions of the Strategic Defence Review and Strategic Security Review. The ATI would also need to strengthen its relationships with the private sector. To ensure the ATI’s success, government will maintain its current level of R&D funding from national security and defence for the next three years, increasing the defence and national security staff embedded within the ATI, and provide refreshed high-quality research problems for the next funding period. This support will be subject to meeting a renewed set of KPI’s attached to this investment.

This would mean that the ATI’s current non-defence activity would need to be reoriented to support this renewed focus and strengthen the UK’s sovereign AI capabilities. In practice, I suggest the Institute’s work adheres to the following principles: (1) delivers cutting-edge AI research of the type that cannot or would not be undertaken by industry or individual universities, (2) delivers tangible impact in support of government missions, (3) facilitates interdisciplinary collaboration and ecosystem building, and (4) forms part of a coherent programme of work that furthers government’s defence, national security and sovereign ambitions. In this capacity, impactful existing work on weather prediction and fundamental AI research would align well, albeit with a renewed emphasis on, and specific linkage to, the UK’s defence, national security, and sovereign capability.

Regarding the ATI’s existing community of research software engineers, I would welcome the development of a contracting system for government to direct engineers towards deep tech R&D challenges in support of government’s AI ambitions. As the ATI seizes the opportunity to deliver big impact for government through projects such as this, we envision the institute will develop a clear value proposition with respect to international talent. This will build on government’s commitment to work with the ATI and UKRI to drive progress at the cutting edge, support the government’s missions and attract international talent.

To realise this vision, it is imperative that the ATI’s leadership reflects the institute’s reformed focus. While we acknowledge the success of the current leadership in delivering reform at the institute during a difficult period, careful consideration should be given to the importance of an executive team who possesses a relevant background and sector knowledge to lead this transition. Similarly, the Board of Trustees should be made up of experts aligned with the institute’s renewed purpose.

Regarding funding, I believe the ATI should continue to receive the funding needed to implement reforms and deliver Turing 2.0. That said, I anticipate that delivery of the vision outlined above will catalyse additional investment, both from within government and through related commercial opportunities. This, coupled with cost reductions from ongoing efficiency savings and organisational change, should lead to reduced dependency on EPSRC core funding from 2026/27 onwards. As such, it may be possible to reconsider the institute’s longer-term funding arrangement during the 2026 mid-term review.

I will ask my officials work closely with you in delivering this vision, and in the coming period I encourage continued close working between DSIT, ATI, EPSRC, and relevant stakeholders to ensure a smooth transition while maximising the ATI’s unique value-add.

Thank you for your continued dedication to advancing the institute’s mission at this crucial juncture.

Yours sincerely,

Rt Hon Peter Kyle MP

Secretary of State for Science, Innovation and Technology

Jean Innes’ message to Turing staff

Dear all,

As you may have seen in media reporting, as part of our ongoing dialogue with Government, Secretary of State Peter Kyle has written to the Turing setting out his vision for how the institute can build on its existing strengths and best support the Government’s AI ambitions.

This resulted from a recent meeting with the Secretary of State to discuss the AI Opportunities Action Plan, and specifically the Plan’s commitment to “Consider the broader institutional landscape and the full potential of the Alan Turing Institute to drive progress at the cutting edge, support the government’s missions and attract international talent.”

We understand that information presented in the media, quoting selectively from a letter from the Secretary of State, may be unsettling to read. But it is important to recognise that the letter is collegiate, not critical, endorsing the Turing 2.0 approach, noting our crucial role in the UK’s AI transformation, outlining the strength and impact of our existing work and setting out the Secretary of State’s views on areas where the Turing can uniquely add value.

This is to be expected and welcomed in a rapidly moving ecosystem in which the Turing is already evolving, and where we have also seen recent reforms to the AI Security Institute and the creation of a Sovereign AI Unit, the conclusion of the Spending Review, and publication of the Industrial Strategy and Strategic Defence Review.

The Secretary of State recognised the excellent work of our defence and security teams and the vital role we have as a national security asset. He recommended we build on our existing strengths to prioritise our work on defence, national security and sovereign capabilities.

As you know our high-impact missions are increasingly interlinked. As the Secretary of State said, our impactful work on weather forecasting and fundamental research aligns well with his vision, and there is an opportunity to bring all of that work more closely together to support the Government’s AI ambitions and meet the evolving needs of the nation at this critical time.

As you would expect, we will continue to work closely with the Government over the coming weeks and months to support its AI ambitions, developing and deploying new capabilities for the public good, and will continue to share updates as this progresses.

Jean

December 2024 Open Letter from Turing Staff to the Board of Trustees

Dear Board of Trustees,

In strict confidence

URGENT: Letter of no confidence in the Institute’s Executive Leadership Team

We, the staff, have worked tirelessly to lay the foundations for Turing 2.0, despite the continuous uncertainty and challenges we have all faced over a number of years. Throughout 2024, we have raised concerns directly with the Executive Leadership Team (ELT), only to see them ignored, minimised or misdirected. We now find ourselves with no choice but to escalate these concerns to you, the Board of Trustees. It is now imperative that these issues are addressed openly, transparently and urgently by you, the Board, and the lack of any action or even acknowledgement of our concerns makes us question whether they have even been brought to your attention until now.

An opportunity missed

We have worked collaboratively to help define a strategy that had the support of the data science and AI landscape and was critical in securing our long-term funding from EPSRC. This milestone should have offered us the opportunity to realise our full potential with a new period of stability.

When the Board appointed a new CEO, we were optimistic that a fresh perspective would drive the delivery of this strategy. However, the goodwill generated by this strategy, long-term funding, and a new CEO, in Jean Innes, as well as Doug Gurr as Chair, appears to have been squandered in a very short amount of time by the CEO and the ELT she has appointed. Not only are we at risk of missing an opportunity to play a key role in the landscape, we are more concerningly being led into a situation where the viability of the Institute is under question. We outline this below for the Board to take the appropriate action.

Credibility at risk due to lack of scientific direction

Mark Girolami was appointed over three years ago as Chief Scientist with a primary responsibility to "take responsibility for the creation, refresh and delivery of the Institute’s Scientific and Innovation Strategy having regard to ethics and responsible innovation". Despite this fundamental responsibility, the 'grand challenge' approach, championed by the Chief Scientist back in 2021, has made little or no progress. To date, there are no agreed missions or delivery plans for areas under the strategy. Instead, we have had the same vague proclamations delivered, again and again, devoid of any substance.

The scientific leadership we have had has too often been left rudderless; left behind as both the community and the cutting-edge has moved ahead without us, whilst we are publicly and privately criticised for being behind the curve.

We have repeatedly raised concerns about Mark’s leadership, most recently regarding diversity in the recruitment of his scientific leads. Despite promises from ELT, a report on this issue has continually been delayed since spring this year, leaving the Institute’s EDI stated commitments in question. Staff have not forgotten about this and are still expecting the commitment on the externally produced report to be fulfilled.

Moreover, the S&I Directors, who were appointed as the “cornerstones of Turing 2.0”, have seen their roles rendered untenable. It is clear that commitments made to them to be meaningfully part of the Institute's leadership team and to be given funding to fulfil their ambitions have not materialised. Three of these Directors staked their careers to join the Turing; of those, two have left within nine months over concerns over the leadership, and the third has strongly voiced their dissatisfaction.

With Programmes and their teams all slated ‘at risk’ of redundancy, the future of the Turing’s ability to be a credible scientific organisation is now in serious jeopardy. We operate in a broad and rapidly evolving discipline, but in the midst of high institutional uncertainty our talent pool has narrowed alarmingly quickly as colleagues have left, damaging our scientific capability and capacity to deliver against existing contractual commitments.

Furthermore, the Chief Technical Officer – hired on the premise and presumably ability to oversee delivery teams — cannot reasonably step in to direct the scientific direction in the absence of the remaining S&I Directors being disempowered, and there are only so many people in the ‘inner circle’ who would be willing to take up the necessary 'interim' leadership roles.

We call on the Board to step in to provide their support and expertise to the CEO and Chief Scientist that is necessary to restore momentum and confidence in our agreed scientific vision, and to ensure that a clear, actionable plan is put in place with the appropriate leadership — as expected by EPSRC and other stakeholders in the landscape — to get the Institute back on track.

Transformation — delayed, diluted and in disarray

While the Institute’s long-promised transformation seems to have finally commenced, it comes after more than a year of delays, primarily due to the CEO’s failure to make decisions, communicate clearly, or take responsibility for the Institute’s direction. This in itself is a serious cause for concern as the totally unnecessary delay has meant that pressures on budgets are more acute than they needed to be.

Staff now have serious concerns about ELT's capability and capacity to deliver the required changes within the shrinking timeframe. Regular staff survey data shows an alarming and worsening level of trust.

Key aspects of the transformation continue to be stalled or have been severely diluted under the leadership of the current Chief Strategy Officer (CSO), Donna Brown, who has been promoted twice in quick succession despite her inability to drive the transformation forward. Donna’s reluctance to engage directly with staff is of great concern. Furthermore, the lack of transparency, and misrepresentation of decisions and progress have fostered confusion, frustration and mistrust among staff and partners. This is already having obvious and serious legal, financial and reputational implications to the Institute.

Despite the involvement of external consultants and a dedicated transformation team, it’s evident that there is a significant capability gap in delivering the changes required within what remains a relatively small organisation. The primary challenge for the Institute is that it is conducting an alignment process without knowing what it is aligning to: there are no delivery plans and a decreasing number of scientific leaders — both longer-term and otherwise — in place for the areas outlined in the agreed strategy.

Given the critical nature of transforming to Turing 2.0, we urge the Board to step in directly to support the CSO, both in terms of providing clear direction and ensuring the necessary capability is in place to overcome the current barriers. This will likely involve the Board taking a more hands-on role in overseeing progress, holding both the CEO and CSO accountable for specific outcomes, and ensuring that the transformation is adequately supported at every level of the organisation. If this transformation is not urgently revived with clear, committed leadership and the Board's direct intervention, we risk further destabilising the Institute.

Decreasing stakeholder confidence risking financial viability

The sustainability of the Institute is in question, as concerns from partners and funders — ranging from smaller collaborators to high-profile strategic national and international partners — are increasingly going unaddressed. While the Turing has historically been successful in securing funding, there is growing concern among partners about the Institute’s ability to fulfil its commitments.

Contracts signed by the Institute have not been honoured; strategic partners, who were namechecked in the published strategy, now feel excluded from the future; and questions and concerns raised by partners have not been adequately addressed. These relationships, which are crucial for the Institute’s reputation and financial stability, are beginning to show signs of serious and potentially irreparable strain.

Numerous grants have fallen through due to the Institute’s incapacity for making decisions on time and the loss of partners’ trust, impacting our financial viability. A lack of clear and decisive action to resolve these issues risks jeopardising our funding base and long-term financial health.

The Board should step-in to directly and urgently support Nico Guernion as the Director of Partnerships, providing him the confidence to make timely decisions and begin to rebuild trust with our stakeholders and demonstrate the Institute’s capacity to deliver on its promises. Furthermore, given the mixed messages on finances, the Board should intensify the support it is providing to both the CEO and the Director of Finance and Corporate Services on understanding our true financial position, and transparently communicating that to staff.

If these failures continue, the Institute’s financial viability and its ability to secure future funding will be severely compromised. The lack of clear accountability, scientific direction, delivery plans and consistent engagement has left our partners uncertain about the Institute’s direction. Their concerns are becoming more vocal, and relationships that have taken years to cultivate are now at risk.

Risk to retention and staff morale

Staff surveys indicate a catastrophic decline in trust in the leadership, particularly at senior levels. Staff morale and wellbeing has also become a critical concern, with rising levels of stress and burnout across teams — as acknowledged by the data the People team have shared.

This is no coincidence.

The growing sense of disengagement and frustration among staff is directly linked to a lack of accountability and transparency — which are supposedly Turing values — and poor decision-making by ELT. The scale of the problem is clearly beyond the People team’s capability to support staff adequately, with Clare Randall, the Director of People, primarily accountable.

The first phase of the transformation has placed 143 roles at risk and caused consternation and anxiety among everybody else. The statutory process of consultation in relation to redundancies has been poorly managed, raising the risk of legal challenges, and leading to a further decline in productivity, morale, and retention which will further exacerbate the challenge of becoming Turing 2.0. The Institute is nothing without its staff, whose perspectives should be included to improve future decision-making.

We urge the Board to intervene in providing their expertise and capability (for example, on rebuilding the culture) to the CEO and Director of People, as well as clear accountability for wellbeing outcomes, and a comprehensive plan to support staff and retention through this challenging period. Immediate action is necessary to rebuild trust, support staff, and prevent further erosion of morale, productivity and institutional knowledge. If this issue is not urgently addressed, the Institute risks long-term damage to its culture, intellectual capital, and operational effectiveness.

Request for immediate action to protect Turing's future

As outlined above, ELT's actions are jeopardising the viability of the Institute as a whole. Staff have raised serious concerns about our ability to meet the commitments made to EPSRC in relation to our recently secured funding. While ELT continues to assure us that everything is fine, this message is increasingly at odds with the reality we observe across the Institute.

KPIs are at serious risk of being missed; and promises made to EPSRC remain delayed, unfulfilled, or both. We are deeply concerned that it is only a matter of time before EPSRC becomes aware of this discrepancy, and when they do, it will be too late to tackle the challenges we are already facing.

We call on the Board to urgently intervene and hold the CEO and the rest of ELT accountable for their failure to execute the agreed strategy and provide the leadership necessary for the Institute’s success at such a critical time. Before EPSRC, other funders, and regulators are forced to take action themselves, we urge the Board to proactively commission an independent review of the leadership decision-making, management practices, flows of information across the Institute (including to Board) with input from staff at all levels, to ensure that the Turing can recover its momentum and get back on track.

We request that the Board engage directly with staff and partners to rebuild trust and transparency, and that clear, actionable plans are collaboratively developed with staff representation and rigorously implemented to address the ongoing issues.

The Board meeting on the 5th of December should provide an appropriate forum for Trustees to come to a resolution and begin the process of rehabilitation and engagement with staff before the winter break. We strongly believe that without an objective and comprehensive evaluation of the present leadership, the Institute will continue to drift into an untenable position.

We remain deeply committed to the success of the Turing and have demonstrated this through our continued dedication. This letter is written because of this commitment and sent in the best of intentions. We hope the Board will take immediate action to ensure that the Institute can achieve its full potential in the years to come — otherwise it risks a very serious and public failure.

The Alan Turing Institute concerned staff and signatories

Introduction

UK pension funds have become the scapegoat of choice for politicians and investors. The industry delivers poor returns by international standards and has been repeatedly accused of pursuing an overly-cautious allocation strategy that has resulted in a shortage of private capital for innovation.

Last week, the government unveiled its Pensions Review, which brought together a number of long-anticipated reforms. These include mandating all multi-employer defined contribution (DC) pension schemes to offer at least one default arrangement (the investment strategy that members are automatically placed into) with £25 billion in assets under management (AUM) by 2030. It also reiterated an earlier commitment from funds to invest at least 10% of assets in private markets, including at least 5% in UK private assets by 2030. The government has announced that it will legislate to give itself the power to force funds to comply if they miss this target.

There’s plenty in the review to unpick, much of it is sensible.

Defined Contribution (DC) pensions are the default for most UK workers. In these schemes, individuals and employers contribute a percentage of salary into an investment fund. The value of each saver’s pot rises or falls with the performance of the underlying assets.

The UK DC market is unusually fragmented. At the start of 2023, the UK DC landscape had almost 27,000 DC schemes, of which over 25,000 had fewer than 12 members. In Australia, the top five superannuation funds account for over 30% of assets; in the Netherlands, most workers are covered by a handful of large industry-wide schemes. By contrast, small UK schemes lack the scale to operate. Consolidation is likely a good thing.

However, the UK and venture mandates strike me as an attempt to commandeer long-term savings for a short-term bail out of UK capital markets, which risks leaving everyone, bar an army of middlemen, worse off. These mandates are my focus today.

Subscribe now

A good deal for fund managers, not for savers

In the past, I’ve expressed my scepticism about the idea that there is a big gap in the UK venture market that the government needs to help plug. I suspect market size is the real structural problem for European tech and that the government backstopping venture is largely displacement activity.

Luckily, the review’s authors have focused on the benefits for savers of its proposals, namely better returns and lower fees.

In the impact assessment, the review states that ‘a range of evidence suggests scale could deliver over 10 basis points reduction in fees’.

Government clearly considers this to be optimistic, because the worked example in the document only assumes a 6 basis point cut in fees, trimming the annual charge from 0.3% to 0.24%. Over a full 46-year career, this adds £2.5k to the median earner’s pot.

As far as I can tell, however, this calculation doesn’t factor in the higher fees associated with the private assets the government wants pension funds to invest in. Typical venture capital or private equity structures charge 1.5-2% plus 20% carry on gains. Even if only 10% of the fund goes into these assets, the blended fee increases by about 15 basis points, more than double the consolidation saving. If the pension funds decided to downweight venture in favour of other private assets, it could be closer to 10-12 (but still significantly more than 6). While the pension fund rather than savers technically pays these fees, the costs will obviously be passed on.

Bizarrely, the report actually acknowledges that these vehicles “might have higher upfront costs”, but then doesn’t factor them into its analysis. Maybe the government is hoping that scale will help here. It’s true that private equity has been successfully squeezed on fees in recent years, venture fund fees vary comparatively little. Top-tier funds that don’t struggle to raise also know that they’ll be able to hold the line.

The government throws in a 2% uplift to the final pension pot to account for the benefits of diversification (although the number itself seems arbitrary). However, a single 2% bump cannot compensate for bigger fee drag every year for four decades, unless private-market returns consistently outperform public markets (after fees and carry) at a rate unheard of in UK history.

Even if it were possible to chisel a few more basis points off the fees, this still wouldn’t amount to an argument for investing more in private assets. If there is an argument, it’s that higher returns justify the higher fees. But how consistently do we see these higher returns?

Venture is an unforgiving asset class. While the top decile of funds deliver genuinely impressive performance, the median fund … does not. Depending on how charitable you are with your start year, the median UK venture fund has underperformed a passive UK equities ETF by one to three percentage points per year, while requiring investors to forego liquidity. Even these gains are primarily driven by unrealised valuations that may yet never materialise.

As the single biggest investor in UK venture funds, the UK Government (read: the taxpayer) has subsidised this meagre track record to the tune of billions while enriching a caste of fund managers. These subsidy schemes were, in fact, the subject of my first ever post on Chalmermagne.

The limits of scale and stickiness

At this point, the obvious answer might be to allocate to the good funds and not the bad ones, leading to an industry clear out.

UK pension funds are currently pretty unpracticed at allocating to venture – it’s not the same skill as buying regulated utilities or investing in infrastructure projects. Historically, they’ve been unwilling to pay for the salaries required to hire talented investors or have tried to swerve fund fees via direct investment. But even if we make the (doubtful) assumption that UK pension fund managers could become the world’s smartest allocators overnight, there are several obstacles.

Firstly, there is only so much capacity. The top funds are chronically oversubscribed, sometimes by as much as 3x - 5x. Allocations will go to existing investors first. Flagship venture funds may only accept low tens of millions from all new investors combined – a drop in the ocean for a large auto-enrolment pension scheme.

The funds could just grow bigger, but scaling funds beyond a certain size structurally degrades their ability to generate an outsized return. This is because there is a finite number of good opportunities.

To justify a decade of illiquidity, a venture fund traditionally targets a minimum 3x net return. A £500 million fund can hit that if one £10 million deal turns into a £1 billion exit. By contrast a £2 billion fund needs four such deals. This is challenging in the US, with the world’s largest economy and most mature technology ecosystem. It is implausible for a single fund to achieve this in the UK – simply because a smaller market size logically entails a smaller talent pool. For example, just how many sales leaders are there in the UK who can take a company from £5 million ARR to £100 million and beyond?

Big US venture funds like a16z, Thrive, and General Catalyst are already pivoting into different asset classes (like traditional private equity or public equities), precisely because they are struggling to spend the huge piles of money they’ve amassed.

Secondly, stickiness is limited. In venture, roughly 30% of the funds ranked top-quartile in one vintage slip to median or worse in the next. This could be down to team churn, strategy drift, bad luck, or any other number of possible factors. Only the extreme edge consistently beats public markets. In traditional private equity, persistence rates are even lower.

In short, there isn’t a crack team of people we can just hand the money to, even if they had the capacity to take it.

Buy high, sell low

The laws of financial physics aside, this is arguably a generationally awful time to encourage pension funds to pile into venture or and private equity. This is where things get a bit complicated.

Private markets are still living off the fumes of 2021 prices. The bulk of capital raised by venture came in 2019–2021, when interest rates were near zero and valuations hit historic highs. Those funds are now fully invested but largely unexited: their portfolio companies haven’t IPO’d, haven’t raised fresh rounds at lower prices, and haven’t yet been written down. Instead, they remain marked on fund books at inflated valuations. Yet in the secondary market – where investors actually trade these fund stakes – venture portfolios are clearing at sometimes in an excess of 25% below stated net asset value.

This might sound like a great opportunity to buy on the cheap – but it isn’t for pension funds. Under these reforms, schemes need to meet rapid allocation targets. But pension funds don’t usually buy these discounted stakes in funds on the private market directly; instead, they go through pooled access vehicles. This is because defined contribution pension schemes are operationally expected to offer daily-dealing funds, meaning savers can switch or withdraw at any time based on a daily unit price. Private assets don’t fit this model, as they’re illiquid, rarely revalued, and can’t be sold on demand.

A pooled access vehicle is an investment fund that sits between pension schemes and the underlying venture capital or private equity funds. It gathers contributions from multiple schemes, keeps a small cash reserve for day-to-day withdrawals, and spreads the rest across dozens of illiquid investments. Because those investments are not traded every day, the vehicle re-values them monthly. It then converts that figure into a daily unit price so members can see a price and move their money if they wish – even though the published price may lag the level at which the assets could actually be sold.

The pooled access vehicle is not incentivised to deviate significantly from the valuations reported by the venture funds they’re invested in. In a quiet market with few exits, valuations stick, even if market conditions have changed. If a company last raised a round in 2021 and no new financing has occurred, there’s nothing stopping a venture fund from marking it at the same valuation, even if comparable companies have halved in value. So if the fund’s valuation is stale or inflated, the pooled vehicle simply inherits it, and passes it through to DC savers via the daily price.

Unlike DC pension funds, endowments, sovereign wealth funds, and other investors view private market marks as an internal accounting convention rather than a trading price: they do not owe daily liquidity to anyone. A sudden 30% markdown doesn’t flow into a unit price, and they can absorb the hit in multi-year performance figures.

This means that any pension capital moving into these vehicles today risks inheriting that overhang and paying close to full price for assets the market is already repricing downward. Instructing a group of large funds to move into the space in lockstep all but guarantees it, flooding the market and reducing any leverage they might have. For pension savers the timing risk is acute. Funds raised at the end of easy-money cycles (1999-2000, 2006-07) delivered the weakest long-run returns once valuations normalised.

This mandate cushions professional investors in underperforming funds, while transferring the looming markdown to ordinary savers who want cash at retirement. Socialising the downside of the finance industry isn’t fair or efficient.

1) Bid up prices 2) ??? 3) profit

It’s hard to detect a serious theory of change in the review. My best guess from reading is that it’s something like this:

• By pushing every auto-enrolment scheme into at least one £25 billion vehicle, big-fund buying power will shave a few basis points off charges, giving trustees enough bandwidth to run complex, illiquid mandates.

• This will channel billions into new ‘productive’ assets, which will boost returns for savers and unlock £50 billion of fresh capital – half of which would go into the UK.

• The IPOs of these venture-backed firms will in turn fire up London’s equity markets, boosting growth and productivity.

As established above, this will in practice involve paying more for access to a distinctly mixed bag of private assets, while holding out for a speculative 2% improvement in long-term performance.

Meanwhile, the £50 billion headline figure seems charitable. It assumes:

• DC assets will grow at 17% a year (optimistic based on current trends).

• 40% of all current private market exposure pension funds hold is UK-based (currently only true for a minority of funds’ allocations)

• Every fund will hit its allocation target.

Hitting the review’s headline UK goal will mean channeling £5 billion of new money a year for the next five years into companies or projects. At the moment, we have about £2-3 billion going into UK infrastructure schemes and about £600 million going into private capital from UK pensions. Doubling this overnight will inevitably bring distortion. In the meantime, funds will need to secure another £25 billion of allocation internationally, competing against peers, endowments, and sovereign wealth funds to get into the best deals.

It’s also unclear why these listings would boost London’s equity markets. Even if we did create a successful new generation of UK start-ups, there’s nothing stopping them listing in the US, with its deeper capital markets. I will confess to being personally indifferent to this question. The health of the London Stock Exchange makes little difference to the performance of the wider UK economy. Meanwhile, countries like Israel manage to have successful technology ecosystems, even though many of their successful companies list on Nasdaq. Either way, the government just … assumes this will happen.

It’s tempting to look at the examples of Canada and Australia and to say that we’re bringing the UK in line with international best practice. The difference is, these funds commit to venture because their boards believe the risk/return is attractive, and they spread those bets across Silicon Valley, Israel, China and Europe. They aren’t responding to a government quota anchored in a single, mid-sized venture ecosystem that forces buying even when pricing or deal quality is poor. This is the opposite of how successful overseas funds behave!

Closing thoughts

Behind both the review and successive governments’ haranguing of the pensions industry it’s hard to miss the air of desperation. Amid dismal levels of domestic investment, shrinking capital markets, and chronically weak domestic productivity, the Treasury has decided to conscript the pensions pool as an emergency ballast.

Government has fixated on venture as a potential public policy lever for a while, hoping that with enough taxpayers’ money, it can create a cargo cult Silicon Valley. At the end of the day, venture capital funds are as much a part of the finance industry as any other form of investment vehicle. They aren’t uniquely patriotic, productive, or free of vested interests. Middling fund managers are middling, regardless of whether they buy stakes in start-ups, public companies, or trade coffee derivatives.

As ever, the government appears desperate to avoid confronting any of the tax, listing or planning reforms that make private capital hesitant, instead reaching for the sugar hit of other people’s money.

Subscribe now

Disclaimer: These are my views and my views only. They are not the views of my employer, any pension fund managers, or trades of coffee derivatives. Do not take financial advice from strangers’ Substacks. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

The impulse behind Chalmermagne was the sense that in technology and markets, a lot of bad assumptions go unchallenged, institutions frequently underperform, and we fail to learn the lessons of history.

In that spirit, alongside this Substack, I’ve decided to start a podcast with my friend Sinéad O’Sullivan. Sinéad is a defence economist and multi-asset investor. Her experience spans Harvard Business School’s Institute for Strategy and Competitiveness, the Illinois Institute of Technology, and the European Space Policy Institute, amongst much else. She also did a stint as an elite pastry chef.

We’ve recorded some initial episodes on de-dollarisation, unconventional recession indicators, and India/Pakistan (which we frustratingly finished minutes before the first missiles started flying). Podcast production is new for us both, so if we get any listeners, we’ll continue to refine the editing and the production values.

You can find us on Substack, Spotify, and YouTube, and I’m sure more in time.

Hope you enjoy!

Subscribe to the podcast

In September 1958, the first shots were fired in anger in the North Atlantic since World War II. This wasn’t a tense Cold War stand-off between the US and the USSR. Instead, it was an Icelandic patrol boat firing a blank at a British trawler in the first of three ‘Cod Wars’ between the two countries.

Between 1958-61, 1972-73, and 1975-76, the UK and Iceland engaged in a series of confrontations over fishing rights. In each one of these, the UK, despite its overwhelming military and geopolitical heft, was beaten decisively.

The Cod Wars were big news at the time in both the UK and Iceland, but are no longer discussed frequently outside the context of fisheries policy. These sorts of discrete asymmetric conflicts, however, interest me. They showcase how countries view their standing in the world, their strategic assumptions, and the limitations of doctrine. The Cod Wars specifically, however, also embody a series of common flaws in the UK approach to a range of policy and strategy questions in 2025. In this piece, as a welcome break from UK metascience, I retell the story and cover some of the lessons.

Subscribe now

What were the Cod Wars?

The prequel

From the 1890s onward, British trawlers, primarily from Hull and Grimsby, began to fish off the coast of Iceland. This distant fishing trade became central to the economy of a number of coastal towns over the ensuing decades.

In this era, as the undisputed master of the high seas, Britain set the commonly accepted standard that territorial waters extended three miles beyond a country’s shoreline. The UK saw this limit as critical. It allowed Royal Navy vessels to sail close to foreign shores for military and intelligence operations and enabled the unrestricted sea lanes that facilitated British trade. It was the position of the British government for the nineteenth and the early twentieth century that it would fight a war to uphold the principle.

Danish-ruled Iceland attempted to extend this limit to four miles at the end of the 19th century in order to preserve more of its fish stocks for its own citizens. In response, Royal Navy ships paid Reykjavik two friendly visits over 1896-97. This led to the Danes reconsidering and formalising the three-mile limit in a 1901 treaty. This was the status quo for the next five decades.

In 1951, after the International Court of Justice ruled in Norway’s favour on a four-mile limit, Iceland tried again. British trawlermen and their unions were incensed and the British government tacitly supported their decision to prevent Icelandic fish from being unloaded at UK ports. The blow to Iceland was softened by the US and its allies, along with the USSR, stepping in to buy Icelandic fish. Following a 1956 negotiation with the Organization for European Economic Cooperation, the UK reluctantly backed down and accepted the four-mile limit.

The peace didn’t hold for long. In September 1958, Iceland unilaterally expanded its limit to 12 miles, following a series of inconclusive negotiations. The Icelanders declared that any British trawler sailing within these limits would be detained. The UK government dispatched the Royal Navy to escort its fishermen. On paper, this should have been a rout. While diminished from its imperial peak, the UK still wielded one of the world’s foremost navies, while Iceland had a handful of largely unarmed patrol boats. But it was not to be.

It’s war

On 2 September, the first skirmishes broke out. Grimsby trawler Northern Foam was intercepted by two Icelandic patrol boats and signalled to British frigate HMS Eastbourne for assistance. Using a loudhailer, Commodore Anderson for the UK attempted to defuse the situation, calling out to his Icelandic counterpart Captain Kristófersson: “Kris, Kris, this is bloody daft! I’m coming over to talk to you”. After a friendly debate about the rights and wrongs of the matter, the Eastbourne brought the Icelandic raiding party aboard ‘as guests’. An attempt to detain another Grimsby vessel was repelled by the crew, wielding boat hooks, poles, and an axe.

Andrew Gilchrist, the British Ambassador to Iceland, combined an understanding of Iceland’s strength of feeling with irritation at the manner in which they were treating British sailors. At noon on 2 September, after the first skirmishes had been reported on the radio, a crowd of a few hundred protesters, primarily made up of drunken teenagers, had gathered outside his residence. Gilchrist had invited a group of foreign journalists to observe the hostility of the Icelanders first-hand. After the crowd failed to do more than chant a few slogans, Gilchrist blared ‘The Barren Rocks of Aden’, a jaunty pipes and drum military march, on the gramophone towards the crowd. Amid cries of ‘Vikings never give up’, members of the crowd stormed the garden, smashing glass.

Things would only get worse. Unwilling to take Britain’s ‘victory’ on the first day of the Cod Wars lying down, the Icelanders began to fight back. Heavily-plated Icelandic patrol boats began to ram both British trawlers and their naval escorts. A number of tense stand-offs ensued, where Icelandic patrol boats would fire blanks at trawlers. The trawlers would ignore them and the navy would show up and threaten to open fire. Tempers became heated, with one naval officer shouting at an Icelandic patrol boat over the loudhailer, “if you try to ram me, I will blow you out of the water”.

Anderson warned the Admiralty, “I appreciate the political problems but urge a face saving interim solution to be found quickly for both countries if only to save the needless loss of life.” On one occasion, a trawler and an Icelandic gunboat pelted each other with dried cod and rotten potatoes.

While Britain overwhelmingly outgunned the Icelanders - it suffered from two weaknesses.

Firstly, the optics of the Royal Navy showing up to push around a tiny, newly-independent country played dreadfully. While other NATO nations and international courts believed Iceland to be in the wrong, there was widespread sympathy for the Icelandic underdog, even among the British public. Had members of the Icelandic coast guard been killed in clashes, the mood could well have turned poisonous.

Secondly, while the Navy was largely successful in allowing British trawlers to continue fishing - it was a costly affair. The Icelandic patrol boats were operating in the immediate vicinity of their home base, while British frigates had a journey of more than a thousand miles back to port. Despite their superior firepower, British frigates were lightly armoured vessels primarily used to chase Soviet submarines, making them vulnerable to Icelandic ramming attacks.

Mounting costs and rising international opprobrium might have been enough to help Iceland win, but it had a trump card. Two years after joining NATO in 1949, Iceland had allowed the US to establish a military base near Keflavik, on the southwestern tip of the island. This base served as a crucial node in American nuclear strategy, variously acting as a base for long-range bombers and a listening post for submarines. Soviet nuclear submarines had to pass through relatively shallow straits between the north coast of Scotland and Greenland on their way home, making Keflavik the perfect spot for sonar installations to monitor them.

In all three of the Cod Wars, Iceland threatened to leave NATO and cancel its bilateral defence agreement with the US. In the Third Cod War, it went as far as breaking off diplomatic relations with the UK.

These threats were given credence by UK and US fears about communist influence in Iceland. The two countries had (wrongly) suspected that anti-British sentiment had been whipped up by the far-left, ignoring indications that it was an issue of cross-party consensus. Iceland’s fish sales to the USSR, combined with the growing Soviet embassy in Reykjavik, which had a staff contingent double the size of Washington’s representation, did little to allay these fears.

In each one of the Cod Wars, Iceland’s suggestion that it might leave NATO and close the airbase helped prompt a British climbdown. There isn’t clear evidence that the US instructed the UK to back down in the face of this threat; it’s perhaps more damaging for British pride that they didn’t even need to.

The first Cod War ended with total British capitulation in 1961, with the UK accepting the 12 mile limit in exchange for some interim fishing rights.

The second and third Cod Wars followed somewhat similar trajectories, only with greater vehemence.

In the second Cod War of 1972-73, the Icelandic coast guard began to deploy net cutting devices, which increasingly made it impossible for unescorted British trawlers to fish.

To avoid a repeat of the first Cod War, the UK initially used unarmed tugboats to interpose themselves between the trawlers and the Icelandic patrol vessels. This idea had been floated half-jokingly in a meeting at the Ministry of Agriculture, Fisheries, and Food only to be subsequently adopted as policy, with the backing of the Ministry of Defence and the Foreign Office. In a bureaucratic confusion over shipping registrations, Statesman, the first such vessel to be deployed, sailed under a Liberian flag.

Unsurprisingly, this did not get the job done and the Navy was summoned back. This phase of the conflict was less gentlemanly. There were mass protests in Reykjavik, live rounds blew a hole in the side of a British trawler, while one member of the Icelandic coast guard was killed in a ramming incident, when water coated the electrical equipment he was operating.

After NATO brokered talks, the Navy left Icelandic waters, while British trawlers played Rule Britannia and The Party’s Over on their radios. The UK accepted Iceland’s 50 mile limit.

The third and final Cod War followed a similar trajectory to the second, only with Iceland pushing for a 200 mile limit. This saw the most shots fired and the greatest number of ramming attacks, with 15 British frigates damaged. HMS Eastbourne, the vessel from the UK’s triumphant first skirmish in 1958, incurred such bad structural damage that it was never judged seaworthy again.

However, the UK was on much weaker legal footing by 1975. The ongoing United Nations Conference on the Law of the Sea (UNCLOS III) was already moving toward recognising 200-mile exclusive economic zones, while expansive limits were increasingly becoming a global norm. In June 1976, after a mere seven months, the UK capitulated.

What went wrong?

One of the most striking features of the Cod Wars was their similarity. Over a span of 18 years, the British government repeatedly deployed the same set of tactics to no avail. As former ambassador to Iceland Andrew Gilchrist put it:

“If some reasonable degree of excuse and explanation can be offered for the British government’s actions in … 1958, surely it passes comprehension that when confronted by an identical problem in 1972 and 1975, the government should have had recourse to the same measures which had proved so ineffective and counterproductive on the earlier occasion.”

Britain’s failure to change its approach resulted in worse outcomes. Before the second Cod War, the UK rejected Icelandic proposals that were more generous than the settlement it obtained post-conflict. Meanwhile, Iceland actually hardened its negotiating position over the course of the third Cod War, as growing violence soured domestic opinion.

So how did the UK settle on the same set of failed policies each time?

Failure of imagination

A significant driver was a failure to understand the other side. A 1963 account of the first Cod War observed that there was little interest from outsiders in understanding the Icelanders on their own terms:

“Visitors tend to be interested mainly in unusual characteristics (geysers, glaciers, boiled sheeps’ heads, and the like) or in topics directly relevant to their own country. Americans, for example, often write about the Keflavik base or about the threat posed by Communist strength in politics and the labor movement.”

While fishing was important for the UK economy, it was existential for Iceland. Between World War II and the first Cod War, fish had accounted for more than 87% of Iceland’s exports, rising to 97% in some years. It did not occur to planners in London that simply ramping up the display of force, as they did in the second and third Cod Wars, would not scare their adversaries into submission.

Iceland had secured its independence from Denmark in 1944 and for many Icelanders, the Cod Wars were the next front in their struggle for freedom. They had secured their political independence, but not sovereignty over the country’s most important resource. During the second Cod War, many Icelanders would put bumper stickers on their cars, depicting the map of Iceland encircled by a chain.

This mood was apparent to British officials in Reykjavik. As one noted in 1959:

“The Government] had failed dismally to understand the Icelandic psychology and make the necessary allowances for the sensitivity of a very small country which had so recently gained its independence. … [they] might have applied some of the lessons they had learned the hard way in Ireland, with which there were striking historical and temperamental similarities”.

This psychological failing was clear when Ólafur Thor, Iceland’s industries and fisheries minister visited London in 1952 during the prelude to the first Cod War. As a serving minister and three-time prime minister of a NATO ally, he was affronted that not a single minister found time in their diary to meet with him.

This failure of empathy was fused with a lack of institutional knowledge. Iceland’s coalition politics were little understood in the corridors of power in London. Different parties would jockey for position, attempting to outflank each other. Amid this confusing picture, the UK defaulted to assuming the worst, believing Iceland’s true intention was to exclude all foreign fishermen from its waters.

Fear of decline

In 1849, a mob in Athens attacked the house of David Pacifico, a British subject and former Portuguese consul-general to Greece. Allegedly, the police stood by and failed to break up the crowd, which included the sons of a government minister.

In response, Foreign Secretary Lord Palmerston sent the Royal Navy to blockade Greek ports until compensation was paid. Defending his actions in a debate in the House of Lords, Palmerston said:

"As the Roman, in days of old, held himself free from indignity, when he could say, Civis Romanus sum, so also a British subject, in whatever land he may be, shall feel confident that the watchful eye and the strong arm of England will protect him from injustice and wrong."

This was the golden age of ‘gunboat diplomacy’: a short display of naval power to achieve diplomatic goals. The British Empire, wielding the world’s powerful navy, were the undisputed masters of this art.

100 years later and economically devastated by two world wars, Britannia was receding on all fronts. The 1956 Suez Crisis, where international outcry forced the UK and France to withdraw from an Israeli operation to topple the Egyptian president, had cemented the notion that the UK was incapable of pursuing an independent foreign policy. A year after the fiasco, the defence budget was cut by 10% and Minister of Defence Duncan Sandys came close to scrapping the UK’s entire contingent of aircraft carriers.

As a result, successive UK governments were acutely sensitive to anything that could be portrayed as another symbol of British decline. Being pushed around by a demilitarised country of 200,000 people was one such example.

The military felt this acutely too. In the run up to the third Cod War, the Navy was facing the prospect of severe budget cuts. The aggressiveness of their response was in part a product of its leadership attempting to showcase the service’s usefulness.

Fear of decline blinded decision-makers to how the Cold War context and NATO had changed the game. In this new world, the naval threat was ultimately a paper tiger. The UK could never really escalate beyond firing warning shots without opening fire on an ally. Meanwhile, the UK’s weakening economy could not support a potentially endless naval commitment. If a more determined Iceland simply chose to press on, what was the escalation?

Stakeholderism

Finally, the UK government was pushed around by a very important stakeholder: the fishing industry. The industry had felt let down by the UK’s acceptance of the four-mile limit in 1956 and pressed hard for the government to take a tougher line.

In the run-up to the first Cod War, the Hull Daily Mail and other media allies of the trawlermen compared Iceland to Egypt under Colonel Nasser, drawing a parallel between his nationalisation of the Suez Canal and the Icelanders’ approach to fishing limits.

In 1972, the Deepsea Fishing Industry Committee labelled Icelandic actions ‘piracy’ and lobbied hard for protection. If the Royal Navy did not accompany them, British trawlermen threatened to stop sailing to Iceland altogether. This would have amounted to de facto recognition of Iceland’s new 50-mile limit - something the government was keen to avoid. Austen Laing, Director-General of the British Trawlermen’s Federation, warned that 100,000 jobs were at stake.

As is so often the case, the stakeholders did have a point. The final settlement in the Cod Wars was a significant contributor to the decline of the UK’s distant fishing industry, with the impact disproportionately felt in a small handful of towns. In the 2000s, after a long campaign, the government ended up making compensation payments to over 5,000 former trawlermen for the damage done to their livelihoods. The payments, which averaged at £9,700, were roundly condemned for their stinginess.

Stakeholders are often expert in identifying their challenges, but less good at policy design. By pushing the government to adopt a more belligerent stance, the industry arguably hastened its own demise.

Closing thoughts

The drivers of the three Cod Wars were unique, but the flaws they showcased in Whitehall and Westminster persist to this day. Stakeholders have held successive governments hostage over a range of policy areas, whether it’s housing, energy, or science and technology. Meanwhile, the UK’s defence budget has been so cannibalised by vanity projects and misadventure that it would struggle to deploy three or four ships to Iceland, let alone the dozens employed at the heights of the second Cod War.

While the UK had limited leverage in its scrap with Iceland, it is currently, somewhat inexplicably, being beaten up by another motivated island nation on the world stage. In its negotiations with Mauritius over the future of the Chagos Islands, the UK has made unilateral concession after unilateral concession while being in a seemingly stronger position. Iceland, after all, could ram British frigates. Instead of fearing decline irrationally, we appear to have pivoted to embracing it.

Subscribe now

Disclaimer: These are my views and my views only. They are not the views of my employer, the Icelandic coast guard, or any cod in distant waters. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

A note on sources:

This piece is lighter on links than usual, because I relied heavily on books and journal articles. This is a free Substack, not a work of academic history, so I haven’t footnoted every page reference in the interests of readability. My main sources were:

• Morris Davis, Iceland extends its fisheries limits (Oslo, 1963)

• Jeffrey A Hart, The Anglo-Icelandic Cod War of 1972-1973: a case study of a fishery dispute (Princeton, 1976)

• Jon Th. Thór, ‘The extension of Iceland's fishing limits in 1952 and the British reaction’, Scandinavian Journal of History (1992) DOI: 10.1080/03468759208579227

• Gudni Thorlacius Jóhannesson, ‘How ‘cod war’ came: the origins of the Anglo-Icelandic fisheries dispute, 1958-61’, Historical Research (2004) DOIL: 10.1111/j.1468-2281.2004.00222.x

• Gudni Thorlacius Jóhannesson, Troubled Waters. Cod War, Fishing Disputes, and Britain's Fight for the Freedom of the High Seas, 1948-1964, PhD thesis (2004), https://qmro.qmul.ac.uk/xmlui/handle/123456789/1834

• Sverrir Steinsson, ‘The Cod Wars: a re-analysis’, European Security (2016), DOI: 10.1080/09662839.2016.1160376

As a free speech enjoyer, I periodically share a summary of the most interesting critical responses from readers to my work, along with a few reflections.

My posts on null results in policy, cultural explanations of economic performance and European rearmament proved relatively uncontentious, but my account of what went wrong with the Alan Turing Institute, the UK’s national AI and data science institute, gained more traction than I’d expected.

While the majority of responses were supportive, not everyone was happy. Life is too short and I’m not petty enough to respond to every point, but there are a few things I’d like to address.

Separately, a few days after the piece was published, Jean Innes and Doug Gurr, the CEO and Chair of the Alan Turing Institute respectively, gave an interview to the FT about their planned ‘revamp’. I have no idea if this was done consciously in response to this piece or if the timing was coincidental, but I’ll share a few reflections on it at the end, along with some thoughts about what I’d do if I were in government.

Subscribe now

Is Tonty Blair behind this?

A common theme across responses was an attempt to read something sinister into my motivations, along with a peculiar fixation on the Tony Blair Institute (mentioned once in a 4,000 word piece).

Joanna Bryson, a professor of ethics and technology I referenced in the piece, was among those with questions. Her full comment is quite long and available in full here, but the bit that jumped out at me:

If British work on AI ethics was marginalised in 2023 it might be because Brexit was excluded them from working on the EU's AI Act, or because they were panicking about the financial collapse of the country, making enormous portions of their staff redundant, and so forth.

Tony Blair and the AI Safety Institutes seem overly captivated by the same West Coast Accelerationists that are dismantling the US state, committing crimes against humanity, and merging US foreign policy with Putin's. I know there are some extremely smart, yet weirdly ignorant people at both. It's not easy to keep up with the spiral of interlocking concerns – economy, security, technology, governance – but between them the TBI & ASI have enough money and brains I really think they ought to be able to come up with something better if they stopped giving too much credence to charletons whose primary drives seem to be ending restrictions on money laundering – well, that, and validating their horrific lack of concern for the lives of other people.

Meanwhile, back to the universities. They are one of the UK's leading sectors, or were. Maybe still are and the whole UK is crashing, IDK. But we were until recently leading the world along with the US and Switzerland in terms of quantity in the top 100 globally per capita. I'm presently working in Germany because it's still in the EU, but everyone knows German universities are weaker because they have to focus so much on teaching – the research € go to the Max Planks. As far as I can tell, having researchers mixed in with undergraduates, postgraduates, postdocs, and professors has been a winning Anglo-American strategy for a few centuries. I was honestly astonished that the US and UK didn't stand out more in your Figure 1. This is part of what makes me wonder if this is hatchet job, and what the motivation for the piece is.

Bryson then followed this up with posts on LinkedIn and BlueSky that made similar points, including one that asked why I had been ‘set’ on the Turing.

An anon also weighed in to accuse me of writing a ‘hatchet job’:

This is quite entertaining as a hatchet job, but is not a balanced or well-enough informed piece. For every criticism mentioned here you could equally find an opposing view. The EPSRC was not 'politely scathing', but made some pretty standard recommendations for improvements to governance, alongside lots of praise for certain of Turing's research and community-building activities.

Likewise there is a fundamental misunderstanding here about the Institute's role – as there was in the Tony Blair (techno-libertarian) Institute report that criticised the Turing for not having produced 'UKGPT': it isn't, and never was meant to be, a research lab, and has a budget thousands of times smaller than OpenAI (which is losing vast sums of money each day). The Turing's role was to convene university and other research, which it has done very effectively, while building an outstanding team of research software engineers.

To disappoint members of the reality-adjacent community, I was not ‘set’ on the Turing by a shadowy west coast ‘techno-libertarian’ cabal and I don’t … advocate money-laundering? The editorial calendar of this Substack is determined entirely by my personal whim. One of the most common themes since I started it last October has been institutional failure, across technology, finance, and defence. The Turing is a really good example of it.

I quoted a Tony Blair Institute report in the introduction, because it articulated some good criticisms of the institute, not because of some malign intent. The TBI report also doesn’t, as the anon contends, criticise the Turing for not creating a ‘UKGPT’. Similarly, unless the anon has a long track record of working at dysfunctional institutions, I’m not sure the EPSRC (the Turing’s main funder) requesting ‘urgent’ changes to financial reporting and an entire new constitution constitute “standard recommendations for improvements to governance”, but we may have to agree to disagree…

A bizarre comment on motivations came from ‘Jay’:

Really interesting to see the framing here. "Caring about the environment" is political but funding defense research is apolitical and "just makes sense." Huge bias in this post.

This misrepresents the section on defence. The point I was making was that a national institute is well-positioned to work with the government on more sensitive issues. Anyone at a start-up or a private lab can attest to the endless world of clearances that you have to navigate before you can even get through the door. This is not the case for private actors trying to work on applications of AI to environmental challenges. No good faith reading can interpret this post as saying that ‘caring about the environment’ is ideologically suspect.

Gopal Ramchurn, who runs Responsible AI, was interested in why I hadn’t given the AISI and ARIA the same treatment as the Turing and RAI:

I think many in the AI community would agree the ATI could do more. On your evaluation of other UK AI initiatives, I see you've missed an evaluation of ARIA and AISI in your report. Would be good to get your thoughts how they are doing as they have similar budgets to Turing.

I don’t think either of these organisations is above criticism, and if I had any reason to believe they weren’t accomplishing their mission or were being mismanaged, I’d consider writing about them. I think the comparison is unhelpful in this context, however. ARIA is a new organisation and its entire raison d’être is high risk, high reward research. If ARIA had notched up a load of big wins this early, it would suggest that the team wasn’t aiming high enough.

Meanwhile, the point of the AI Security Institute is to evaluate the security of frontier models. Big labs are open about the fact they are working with them on exactly this. ‘Organisation fulfills intended function’ didn’t strike me as a hugely interesting story, even if it is an increasingly rare one.

DeepSeeking justice

Much ink was spilled on my criticism of Mike Wooldridge, who ran foundational AI research at the Turing between 2022 and 2024. I later updated my piece to reflect the fact he left the institute last year, as Turing had listed him as a current employee until mid last-week.

Simon Thompson, a former industry fellow at the Turing, led the charge on this point. His full comment and follow-up replies are worth reading, in part, because beyond Mike and DeepSeek, we probably agree on quite a lot. It’s clear that he sunk a great deal of personal and professional effort into the institute. I respect his openness and honesty.

For a start the weird ad-hominem attacks on Mike Wooldridge. Mike took on his leadership role at the ATI in 2022 when it was (I think) a mess. He is not responsible for the mess, I don't know what he's done about it, but to attribute it to him is just flat wrong. And, yes - as part of his mission to communicate about AI (including doing the Christmas lectures) he has described LLMs as prompt completion - which is exactly what they are. The shock was twin fold, that massive computation and data would allow prompt completion as good and useful as it is and that someone would spend $10m on a single training run for 175bn parameter one. These two things show that our understanding of natural language and economics were both radically wrong... Also AI is not a monolithic discipline and Mike is not an NN or NLP researcher, he is a specialist in game theory and multi-agent systems, so this is like criticising a condensed matter physicist for not predicting dark energy. Anyway, take it from me, you've taken aim at the wrong person and it's unpleasant and unfair.

The thing about Deepseek is especially egregious. Of course he hadn't heard of them. Until V2 hardly anyone I know had, they made some interesting models but nothing really significant. V3 changed that. The sponsorship of the CCP was an important part of this change, if the CIA didn't know why should Mike?

To be clear, I’ve never interacted with Mike and bear him no personal animus whatsoever. I also don’t blame him specifically for anything that went wrong at the Turing in the piece. I included some of his quotations because I feel they’re representative of a wider trend in the academic caste that the government has historically turned to for advice.

It is, however, patently untrue that no one had heard of DeepSeek before January 2025. I (by no means an expert on Chinese AI labs) was writing about them in July of last year, covering the efficiency of their models, their use of synthetic data, and ability to outperform OpenAI models on certain metrics. I’m not claiming any unique foresight here! My friend Shakeel was writing about them at the same time and DeepSeek’s model releases were covered in freely available media.

I didn’t include the DeepSeek quote as some kind of cheap shot. If an academic is going to hold a senior position at a national institute or offer advice to policymakers (specifically on LLMs and generative AI), it’s completely fair to scrutinise their comments on the subject. By the same token, as a believer in free speech, I completely respect people’s right to find me doing so egregious.

Even if we did believe that it was fine for senior academics not to have heard of DeepSeek, it still raises an important question about the quality of advice that the government is receiving. If it can’t rely on a national AI institute, exactly who should it be turning to?

The lights are on, but is anyone home?

One of my favourite responses came from Seán Ó hÉigeartaigh, the Director of AI: Futures and Responsibilities at the University of Cambridge. He noted that:

This delivers a real broadside against UK academia towards the end, but I think it's worth considering it in the context of the ongoing brain drain.

When I arrived in Cambridge, Zoubin Ghahramani was the Big Name & influence, and Cambridge was the leading AI university in Europe. Now Zoubin's playing w infinite resources at Google, lot of the top talent in engineering has gone to industry, & the agenda both at Cambridge & in places Cambridge influences like ATI is much more influenced by the academic holdouts who don't 'buy' LLMs/frontier AI.

Hard to know how to break that cycle. Interesting to me that in terms of vibes, feels like Cambridge AI went from quite liking what we do (frontier AI governance) to very not liking it, even as one would think evidence supporting our research directions was mounting not diminishing.

Again, I think it's super important academia has people who take frontier AI seriously. But if your interest is in building cutting-edge systems, lot easier and more fun to do with the resources of industry. So you still get some governance folks, but ever-fewer scientists.

Also having done some unfun management jobs in my time, I reckon "Executive Director of ATI" is probably a candidate for "least fun job in the UK".

I’d considered including the brain drain angle in the original draft, but was conscious that the post was well on its way to being a six-part drama lengthwise. I definitely think there’s been a strong (largely inverse) selection effect in UK academia.

Government has two choices in response to this: fund good research or change who advises it. It appears the government has opted for the latter approach in recent years, which obviously comes with trade-offs. Relying heavily on the Turing and the old AI Council combined the worst of both worlds - not funding academia and relying on it.

Do I feel sorry for the leadership of the ATI? Yes and no. As I document in the piece, I think they were dealt a pretty rotten hand, but I also don’t think they’ve played it well…

Scream if you want to go harsher

One line of criticism I wasn’t anticipating was that I hadn’t been harsh enough. But some anons seemed to think I hadn’t taken the current management to task sufficiently on redundancies and project closures.

One said:

​​It Is a good piece. However, I wish you would have covered more recent context to properly understand the current cuts. For example, the fact that three out of the four science leads that were supposed to make the new strategy thrive have resigned; one of them being vocal about a lack of faith in the institute's leadership and recently hired as research director of DeepMind.

And another:

you are missing crucial information about the current administration that has been tanking the ATI lately. the closure of 1/4 of the projects is the tip of the iceberg. those projects were the only ones alive when the formal layoffs begun. in reality closures started with the creation of the chief scientist position, a sort of tiny dictator in the ATI. he has been strangling projects and research groups left and right by killing partnerships and cancelling grants. Mike Wooldridge is one of many examples of the people he pushed out. a more accurate number of closed projects would be around 2/3. it would be useful for you to talk to external partners who have engaged with the ATI recently to see the damage that this strategy has caused.

These anons seem well-informed and I suspect are current (or newly former) Turing insiders. Unfortunately, I’m not omniscient and I can’t make people talk to me. If someone hasn’t supplied me with information or it’s not available publicly, I can’t cite it. There were lots of bits of third hand gossip or tittle-tattle that came my way while I was researching the piece, but I can’t print claims that I’m unable to stand up.

If you wrote one of these comments - get in touch. While I will always protect your anonymity, I can’t uncritically reprint information when I don’t know who you are and have no way of verifying it. You could be part of the techno-libertarian cabal, after all…

Back to the future: the Turing responds (sort of)

Jean Innes and Doug Gurr’s FT interview was a sorry excuse for an institutional defence. It stopped short of addressing any of the substantive criticisms that the Turing has faced. The headline takeaway was that the institute would adopt a more focused approach, doubling down on health, the environment, and defence as part of a ‘revamp’.

“Britain’s flagship artificial intelligence agency will slash the number of projects it bucks and prioritise work on defence, the environment and health as it seeks to respond to technological advances and criticisms of its record.

This is not a ‘revamp’. This is a restatement of the Turing 2.0 strategy that they adopted in 2023. I referenced this strategy in the piece - it’s the 66-page document with a single mention of LLMs. The only new news is that the institute will be doing so with a reduced headcount.

One also has to question the quality of the institute’s oversight. Gurr, the current chair of the Turing, is also Director of the Natural History Museum and the interim chair of the Competition and Markets Authority, at a time when the CMA is under serious pressure from government.

The FT, to their credit, pushed on this:

Asked it was possible to do the three jobs at once, Gurr joked that he was “unbelievably good at time management”.

This doesn’t strike me as a satisfactory response. I equally wasn’t blown away by his attempt to describe the Alan Turing Institute’s value proposition for potential talent:

“Our proposition is that, while you can get more by going across the road to a large tech company, what you get here is the opportunity to work on some of the most interesting problems that will make a real impact on the world,” Gurr said.

Is Gurr really suggesting that the median researcher at the Turing would be working on more interesting problems than a peer at Google DeepMind? Is the Turing’s work really having more of an impact than AlphaFold?

Reflections

I first became aware of the Alan Turing Institute in 2017 and couldn’t quite figure out what it did, but paid little attention. I took more notice over the course of 2023 and 2024, when friends or acquaintances with otherwise uncorrelated opinions began to make similar points about the institute and its lack of impact.

Researching my piece convinced me that it’s unsalvageable in its current form, even if it were to receive better leadership and oversight. Its remit is too broad and its structure too broken for it to have any realistic prospect of a turnaround.

Admitting failure and starting again will be difficult, but I am not aware of any legal obstacles preventing research councils from redirecting their funding to better ends.

I hope the Alan Turing Institute brand is preserved in some form and I see no reason why it couldn’t be applied to a new body that acts as a more fitting tribute.

I don’t have strong views on the exact role of the new institute, but there’s a case for a cyber or national security focused body, preserving the best of the Turing’s defence work. Alternatively, if the government wants a national institute to have a broader focus, it could bring together leading next generation researchers from different universities under a single institutional umbrella and fund them to focus on open frontier research.

Whatever road they choose to go down, they need to ensure they pick a single focus and stick to it. Building an ‘institute for everything to do with AI and the public sector’, or ‘AI, the five missions and badgers’ is a surefire way to ensure you create Turing 3.0.

This will probably be my last word on the Alan Turing Institute for a while, but I’m open to revisiting this. If you have information you’d like to share in confidence, please get in touch.

Subscribe now

Disclaimer: These views are my views or those of the people leaving comments. Unless they left them ironically or didn’t mean them, in which case, they should put more thought into how they present themselves in public forums. But more importantly, these aren’t the views of my employer or anyone who didn’t leave a comment. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Introduction

The UK’s national AI institute is in crisis. Despite receiving a fresh £100 million funding settlement in 2024, the Alan Turing Institute (ATI) is gearing up for mass redundancies and to cut a quarter of its research projects. Staff are in open revolt.

In 2023, when generative AI fever swept the world, far from taking centre stage, the ATI found itself playing defence. A report from the Tony Blair Institute argued that:

While industry figures have predicted some advances and privately warned of major risks, existing government channels have failed to anticipate the trajectory of progress. For example, neither the key advance of transformers nor its application in LLMs were picked up by advisory mechanisms until ChatGPT was headline news. Even the most recent AI strategies of the Alan Turing Institute, University of Cambridge and UK government make little to no mention of AGI, LLMs or similar issues.

Martin Goodson, then of the Royal Statistical Society, dubbed the ATI “at best irrelevant to the development of modern AI in the UK”.

The criticism hasn’t slowed. Matt Clifford’s AI Opportunities Action Plan recommended that the government should “consider the broad institutional landscape and the full potential of the Alan Turing Institute to drive progress at the cutting edge”. While Clifford was too diplomatic to say the quiet bit out loud, in government speak, this is code for suggesting that the ATI was not fulfilling its intended purpose.

UK Research and Innovation, the UK’s main research and innovation funding body, is running out of patience. Its quinquennial review of the ATI, published in 2024, was politely scathing about the institute’s governance, financial management, and the quality of its most recent strategy (dubbed Turing 2.0).

There aren’t many areas of consensus across the UK’s fractious tech community, but the ATI has come to play an oddly unifying role. From left to right and north to south, there’s a sense that the institute is running out of friends and time.

While the median founder, investor, or civil servant will cheerfully roll their eyes at the mention of the ATI, what’s less understood is why the UK’s national AI institute has proven such a flop, despite spending over quarter of a billion pounds since its inception. To piece this together, I’ve spoken to current and former Turing insiders, figures from the world of research funding, academics, and civil servants.

Subscribe now

Set up to fail?

To understand the ATI, we need to understand the circumstances of its creation. Back in 2014, we had hit the peak of the ‘big data’ hype cycle.

As data storage costs tumbled, big tech companies printed money, and the cloud became the rage - the UK Government decided it wanted a piece of the action. ‘Big data’ had been identified as one of the UK’s ‘Eight Great Technologies in 2013’, unlocking new funds for consultancies to produce pdfs on public sector innovation.

In the 2014 Budget, George Osborne announced that the UK would invest £42 million over five years into a data science institute, dedicated to the memory of Alan Turing. From day one, the goals of the institute were nebulous. The official government announcement stated that the institute would “collaborate and work closely with other e-infrastructure and big-data investments across the UK Research Base”, as well “attracting the best talent and investment from across the globe”.

The task of bringing this ambition to life fell to the Engineering & Physical Sciences Research Council (EPSRC), a powerful public body tasked with allocating money to research and postgraduate degrees. In July, a few months after the Budget announcement, the EPSRC called for expressions of interest from university partners. It would go on to announce Cambridge, Edinburgh, Oxford, UCL and Warwick as its initial partners following a competition to which over 20 universities applied. It has since expanded its university partnerships to a network of over 60.

The government announced that the ATI would have a headquarters in the British Library in London (after turning down a bid from Liverpool), along with “spurs” in businesses and universities. This would “bring benefits to the whole country … including in our great northern cities”, despite no academic institutions from the latter making the cut.

Creating a national data science institute in partnership with some of the UK’s top universities might seem like a logical move. But arguably, this was the root from which all of the ATI’s ills stemmed.

A virtual institute

By outsourcing the institute to the universities, the EPSRC guaranteed that the ATI would always be less than the sum of its parts.

In the words of one senior academic: “The EPSRC wanted to establish a data science institute, but didn’t want to take on the long-term responsibility of funding and governing it … the landscape is filled with institutions that have been offloaded onto higher education”.

Universities contributed research muscle, but if an academic became a ‘Turing fellow’ or led a programme at the institute, they would retain their primary affiliation and would often work no more than a couple of days a week at the ATI.

This made it hard to build a research culture or set of owned capabilities. One ex-ATI staffer described the institute as “a set of sole traders … all self-interested”, who had “little incentive to collaborate”. While they remembered some of their former colleagues contributing, others treated the ATI as “a fancy coffee room”. A current Turing staffer said, “the academics just have no skin in the game”.

A senior research funder adds, “If you’re not aiming to build an institutional culture, then why set up an institution?”

Another source described how “it was conceived as a weird virtual institute”, which meant that “universities bickered with each other and then with the EPSRC. There were too many cooks, no core of people who could deliver stuff, and they [the ATI] never had their own strategy.” To illustrate the many cooks, there are 25 people who sit across the ATI’s three advisory boards, only one of whom works for a frontier AI lab.

There was no bigger source of bickering than money. While the EPSRC stumped up much of the ATI’s money, partner universities were big financial contributors. They expected their money back - with interest.

Each partner university nominated a ‘Turing liaison’ to act as the contact point between their university and the institute. In reality, the liaison director’s job was to maximise the amount of money awarded to their own institution’s research. As our former Turing insider from before puts it, “the original five universities were customers and suppliers”, this “created an obvious conflict of interest”.

If we take 2017 as an example:

The original five universities paid in a million apiece, and made a good return on their investment.

As the ATI’s group of partner universities expanded, the original five were the predominant beneficiaries. While they took a ‘loss’ on the first couple of years of contributions, they soon handsomely profited. In 2019, for example, the five put in £5 million and received back £9.8 million in grants. The new universities in the coalition who had contributed £8 million benefitted from less than £4 million.

It likely doesn’t hurt that the five founding universities hold the majority of seats on the ATI’s board of directors, along with veto rights.

UKRI has grown weary of this setup, with the quinquennial review warning that “there is a clear need for the governance and leadership structure to change fundamentally to reflect a more representative set of stakeholders … the Institute must have an impartial governance arrangement which is fit for this purpose, which follows best practice for similar public investments, and be shown to represent the concerns of the whole community”.

A national institute for hire

How were the universities ‘profiting’?

In part, the ATI acted as a vehicle for shuffling public money from one institution to another. But from day one, the institute had commercial ambitions. The ATI quickly established itself as a taxpayer-subsidised technology consultancy.

As one former ATI staffer put it, the pressure from universities to maximise returns meant that the institute would “go where the money is … it didn’t smell right”.

From 2016, it began striking partnerships with outside organisations. Work the ATI undertook included flight demand forecasting with British Airways, fraud detection in telecoms with Accenture, and improving Android game recommendations for Samsung. Other work the ATI conducted, while not directly sponsored by industry, clearly was designed with commercial applications in mind, such as exploring the correlation of viewers’ emotional journeys with the commercial success of films.

This is not to call into question the merit or rigour of any of this work. But it is reasonable to ask why a national institute was conducting it and how this was building nationally-useful capabilities. Industry could easily commission academics directly or pay consultancies to conduct this work for them.

The ATI was also at an, arguably unfair, advantage in this respect. Thanks to its generous upfront funding and academics’ comparatively low dayrate, the institute had a price advantage versus its commercial rivals. It didn’t hurt that they could, as an academic put it, “hawk the national institute” and “trade off the Alan Turing name”.

Missing the AI boom

In 2017, the ATI added artificial intelligence to its remit at the request of the government. The institute, however, struggled to make an impact in its new role.

Again, the universities bear a significant share of the blame.

Despite the frequent use of the ‘world-leading’ sobriquet by the government, UK universities were not at the frontier of contemporary research on deep learning. Senior British computer science academics’ strengths historically lay in fields like symbolic AI or probabilistic modelling. They initially viewed the work of DeepMind and others as little more than a glitzier version of 1980s work on neural networks. If you look at leading UK university computer science curriculums across 2018-2020, deep learning scarcely features. Senior UK academics were sceptical about how much deep learning could scale and some even forecast an ‘AI winter’.

Michael Wooldridge, a senior Oxford academic and director of Foundational AI Research at the ATI from 2022 to 2024 was among the sceptics.

As late as 2023, he was still playing down recent innovation, dismissing AlphaGo as an artifact of scale. In the same piece, Wooldridge dismissed LLMs as “prompt completion … exactly what your mobile phone does”. In response to the DeepSeek mania in January of this year, he said, “I confess I hadn’t heard of them”. This oversight might have been excusable had it not been made by the most senior AI researcher at a national AI institute. [Update (02/04) : at the time of publication, the Turing Institute listed Mike Wooldridge as a current member of staff. It has since been brought to my attention that he left the ATI in the summer of 2024, with the website being updated to reflect that this morning.]

At the ATI’s flagship annual AI UK conference in 2023, the single talk on generative AI (“ChatGPT: friend or foe?”) was delivered by Gary Marcus, a prominent LLM sceptic who had argued a year earlier that deep learning was hitting a wall.

This dismissiveness towards the latest AI research was fused with, as an ex-staffer puts it, “a left-leaning middle management layer … who wanted to narrate the public conversation”. As a result, the ATI positioned itself as “the police officer, not the innovator”.

This is reflected by the institute’s output from 2018 onwards, which skewed heavily towards a specific ideological interpretation of responsibility and ethics.

There is, of course, nothing wrong with promoting the responsible use of AI or ethical development practices. Problems like bias are well-documented and have caused real-world harm. The ATI’s work, however, tended to focus less on building technical solutions and more towards the banal restatement of well-known problems (e.g. the risks of facial recognition in policing, the importance of data protection), unremarkable surveys of public opinion (people think detecting cancer is a good thing, but are nervous about bombing people), or highly politicised content that seems better suited to a liberal arts college.

A current member of staff observed that “the public policy programmes have some of the wooliest outputs imaginable”.

A drinking game based on mentions of ‘stakeholders’ or throwaway references to the need to ‘consult civil society’ would prove fatal.

This cultural bias was also self-fulfilling. As another former insider put it, “many technical academics just got fed up and walked away”. Another bemoaned how the “institutional capture by sociology is total”.

A national institute could, of course, make practical contributions on these questions. The Toronto-based Vector Institute, for example, built an open source tool that not only classifies biased training data, but also debiases it. That said, the ATI did build a closed source bias classifier for … its customers strategic partners at Accenture.

The ATI’s decision essentially to ignore much of the cutting edge work coming out of DeepMind and US labs meant that its leadership was asleep at the wheel as the generative AI boom got underway.

Until early 2023, the ATI’s output did not mention language models at all. In response to criticism, the ATI’s most senior researchers argued that LLMs had caught the whole world by surprise, so it was unfair to single them out.

While these models may not have entered the public consciousness pre-ChatGPT, it is unserious to suggest that they were somehow secret. Nathan Benaich and Ian Hogarth covered them in the 2019 installment of their State of AI Report.1 The Report has been widely read in AI circles for over half a decade. OpenAI’s initial decision not to release GPT-2 earlier that year was covered in UK national newspapers. What the ATI dubbed hindsight, others could term dereliction of duty.

When the government did need to call on the ATI for help, they were frequently underwhelmed. A former government insider who requested the ATI’s help on Covid-related technical analysis was horrified when the institute took 18 months to complete the work. Based on their own technical background, they believe a competent team could have managed it in one. By the time they received the output, the pandemic was over and it was no longer useful. They consistently found the ATI’s leadership more interested in the amount of money they could charge than in how to be the best possible partner.

This is a wider disease in the UK ecosystem. The government’s (now defunct) AI Council, primarily staffed by the great and good of UK academia, spilt much ink on the subject of responsibility, but failed to include LLMs in their 2021 roadmap. One academic familiar with the work of the council said that, “On some level, these people still don’t believe AI is real. They were telling the government that there was nothing novel about the transformer [the architecture that underpins LLMs] and that they should be focused on regulating upstream applications”.

There is little sign that UK academia has learnt its lesson. In 2023, UKRI staked Responsible AI (RAI) with £31 million of public money.

RAI was born after the new Department for Science Innovation and Technology leant on UKRI to free up money, so that it could showcase investment in the UK’s new ‘Five Critical Technologies’ - the government’s latest set of priorities for tech policy. UKRI scraped together a £250 million Technology Missions Fund from old budgets, but struggled to find investable projects.

UKRI turned to the universities for help and a coalition, led by Southampton, conceived of Responsible AI. RAI, the Technology Missions Fund’s flagship AI investment, has gone on to fund a series of Turing-esque projects, including yet more responsibility guidelines, the use of music to keep tired drivers awake, and whether citizens should have personal carbon budgets. RAI’s first annual report lists six achievements, two of which boil down to hosting roundtables and presentations. The purpose of the Technology Missions Fund was to drive the development and adoption of new technology, not to, as RAI put it, “connect the ecosystem”.

As ChatGPT stunned and terrified Whitehall in equal measure over the course of 2023, the ATI found itself on the defensive. But its defence frequently underscored the problem. Then director and chief executive, Adrian Smith said, “We have delivered numerous briefings to government officials, hosted training and contributed to several roundtables on this topic, as well as offering commentary through media and broadcast interviews.” The Turing 2.0 relaunch strategy from the same year contains a single reference to LLMs across its 66 pages.

There are welcome signs that academia’s grip is being broken. Dissatisfied with the advice they received from the universities, the government established the UK AI Safety Institute (recently renamed to the ‘Security’ Institute) in November 2023, to bring it closer to the work conducted in frontier industry labs.

As the highest levels of government grapple with the interdisciplinary risks, such as biosecurity, that have risen from a rapid acceleration in capabilities, the universities remain unrepentant. If anything, senior academics have prided themselves on their opposition to the new direction and have vented their frustration at losing their place in the Number 10 rolodex.

Dame Wendy Hall, a senior computer scientist and former AI Council Chair turned RAI leadership team member, has condemned the “tech bro takeover” of AI policy and accused the government of ignoring cutting-edge work in universities. Professor Joanna Bryson, co-author of the UK’s first national-level AI ethics policy in 2011, recently derided AI safety as “transhumanist nonsense”. Neil Lawrence, who holds a DeepMind-sponsored professorship, slated the UK government’s “appalling” discussion of AGI and accused it of celebrating big tech companies for “solving problems that no one ever knew they had”.

It is, of course, possible to object to specific claims made in AI safety research or to disagree with researchers concerned about existential risk. But the response of academia’s old guard has been to ignore the substance of this work and instead to designate the entire field as ideologically suspect.

The ATI itself doesn’t seem to have learned either. Its 2025 AI UK conference didn’t feature a single researcher from a frontier research lab this year.

What could have been?

Amidst a maelstrom of stakeholder wheelspinning, one area of strength stands out in the ATI’s work: defence and security. While sitting under the ATI umbrella, the defence and security programme has its own team, research agenda, and operates independently. Thanks to its ties to industry and the intelligence agencies, it’s also the least ‘academic’ part of the ATI.

Its work includes directly supporting the application of technology to national security problems, AI research for defence, and partnering with allies in the US and Singapore.

The most publicly visible manifestation of this work is the Centre for Emerging Technology and Security (CETaS). Driven by a clearer sense of mission, CETaS has built up a degree of community respect that the ATI proper has not. A security researcher at Google DeepMind described them as “great… the best bit of Turing”. An academic with experience of working with the national security community said that “they just quietly get on with it”.

Their output consistently engages with frontier research, provides specific and differentiated advice to industry (e.g. on the need to integrate the UK and Korean semiconductor industries), and empirically analyses popular memes (e.g. the (lack of) impact of AI-enabled disinformation on elections) as opposed to parroting them.

CETaS also … just makes sense. There are obvious advantages a public body has when it comes to working with the national security community. When it comes to broader ‘ethics’ work, which any number of think tanks or advocacy groups could do, the relevance is less apparent. Similarly, the intersection of AI and health, another ATI ‘mission’, more logically sits in universities with links to hospitals.

If the government were to re-evaluate the ATI, there would be a strong case for preserving CETaS, potentially as a sister organisation to the AI Security Institute.

A system failure

The story of the ATI is, in many ways, the story of the UK’s approach to technology.

Firstly, drift. The UK has chopped and changed its approach to technology repeatedly, choosing seven different sets of priority technologies between 2012 and 2023. Government has variously championed the tech industry as a source of jobs, a vehicle for exports, a means of fixing public services, and a way of expanding the UK’s soft power. These are all legitimate goals, but half-heartedly attempting all of them over a decade is a surefire means of accomplishing relatively little.

Connected to this is our second challenge: everythingism. My friend Joe Hill described this aptly as “the belief that every government policy can be about every other government policy, and that there are no real costs to doing that”. This results in policymakers loading costs onto existing projects, at the expense of efficiency and prioritisation.

As the ATI’s original goals were so vague, it was a prime target. Even before the ATI was up and running, the government announced that it would also be a body responsible for allocating funding for fintech projects. It then had a data ethics group bolted onto it as a result of a 2016 select committee report. As one ex-insider put it, “there was never a superordinate goal”.

Finally, the perils of the UK government’s dependence on the country’s universities for research. The UK has historically channelled 80% of its non-business R&D through universities, versus 40-60% for many peer nations.

This is the product of policy choices made in the 1980s onwards, which prioritised funding research in universities over public sector research establishments or government labs. This started under the Conservatives, but continued under Labour with the first Blair Government’s decision to close and partly privatise the Defence Evaluation and Research Agency (then the UK’s largest government R&D body).

The other driver is the 2002 introduction of the full economic costing model (fEC) for research funding. When a research council awards a grant to a university, they do not have to fund it in full. They typically only cover 80% of the fEC, while universities pay up the rest. In practice, research councils will often try to drive their share even lower.

Unlike other institutions, universities are able to cross-subsidise research costs from other income sources, such as international student fees. They also benefit from an annual £2 billion block grant that helps them make up any shortage. This system favours universities over everyone else and incentivises the worst kinds of grant-chasing.

It’s perhaps why some of the more interesting institutes and labs attached to universities are the product of philanthropy, rather than the usual cycle of research council allocations. The Gatsby Unit at UCL, where Demis Hassabis completed his PhD, was funded by David Sainsbury. The Peter Bennett Foundation funds institutes at Cambridge, Oxford, and Sussex. A source in the research funding world put it more bluntly: “Rich people can defy university shitness… the Turing is the product of academic capture and the professors that run research councils.”

In essence, the universities ended up in charge of the ATI, because government policy ensured that there was no one else with the necessary scale and resources to take on the responsibility. It just so happened that they were also uniquely ill-suited to the task. It’s not a coincidence that the most functional arm of the ATI was the one with the least academic involvement.

The most common theme of Chalmermagne posts so far has been institutional failure and the legacy debts it creates. If you look across reams of UK policy, whether it’s in technology, defence, or finance, you see the same thing: a penchant for symbolic announcements, a bias towards outsourcing to ‘stakeholders’, and a reluctance to kill off bad policy long after any reasonable trial period.

Considering how the UK treated Alan Turing while he was alive, he deserved a better institute to honour his memory. But then again, government is slow to learn from its mistakes.

Update (02/04/2025): at the time of writing, Mike Wooldridge was listed on the Alan Turing Institute’s website as a serving member of staff. This morning, the Turing updated their website to clarify that he has left the Institute. I have amended the piece to reflect this.

Subscribe now

Disclaimer: These are my views and my views only. They are not the views of my employer, any stakeholders, or civil society organisations. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Introduction

While studying history a decade ago, John Brewer’s The Sinews of Power left an impression on me. Brewer argued that Britain's rise as a global power in the 18th century had little to do with limited government or ancient English liberties. It instead resulted from its development of an efficient "fiscal-military state" capable of effectively raising tax revenue and managing the national debt to fund warfare. This revolutionary state apparatus featured professional bureaucracy, sophisticated financial markets, and a system of long-term public borrowing that allowed Britain to mobilise resources more effectively than its continental rivals, despite its smaller population.

This has stuck in my mind as European governments reckon with a US administration that seems content to abandon its place in the world and engage in the mafia-like extortion of its long-standing allies. We’ve seen European leaders rush to express their solidarity for the embattled Ukrainian government, while governments have pledged to revisit the question of defence spending. European Commissioner Ursula von der Leyen has spoken of freeing up €800 billion for defence investment, although the maths behind this number remains unclear.

But defence is messy. The sinews of power in the 21st century are slightly more complicated than they were in the 18th, but many of the underlying principles are the same. Three hundred years ago, effective defence required good institutions, markets and physical infrastructure. Spending on the military was only one - very important - piece of the puzzle. I’ve written before about Britain’s hollowed out shopfront military; this time, I take a wider European perspective, looking beyond just equipment to cover some of the other sinews. The picture isn’t cheerier.

Subscribe now

I’m afraid there is no money

Defence is expensive and many of Europe’s largest economies are broke.

France and the UK, historically been among the world’s biggest spenders on defence, are out of fiscal headroom. Poland, which is comparatively economically healthy, is already ramping up defence spending. The Baltic states can afford to spend a greater share of their GDP on defence (and are already doing so), their smaller economies mean that the money will only stretch so far.

While most NATO members are now set to hit the 2% spending target, the prospect of a significant US pullback means that this is no longer sufficient. Bloomberg Economics has forecast that European governments could face a defence bill of $3.1 trillion over the next 10 years. This assumes European nations will move towards spending 3.5% of GDP on defence and that they will have to cover the cost of maintaining a Ukraine peace-keeping force and post-war Ukrainian economic reconstruction.

It could be that these Bloomberg spending figures are on the conservative side. These numbers might be enough for Europe to bulk up its armed forces and equip them, but there are certain capabilities that European militaries have essentially never provided for themselves. For example, European states have never fielded an independent ballistic missile defence capability. They lack the necessary radar equipment, while the SM-3 interceptors we currently rely on for upper-tier missile defence sit on US Aegis destroyers. Andrius Kubiliu, the European Commissioner for Defence and Space has previously said that a missile shield could cost Europe €500 billion.

This isn’t the only area of dependence. European governments are also dependent on US multistatic sonobuoys to listen out for submarines and rely heavily on US nuclear attack submarines to trail Russian submarines. The loss of this capability would leave crucial undersea cables undefended.

Is there a magic money tree?

So, how are European governments going to come anywhere close to paying for this?

Conventional measures like reallocating spending or hiking taxes will make up some of the gap, but this will prove politically difficult beyond a certain point. The UK might be able to increase defence spending from 2.3% of GDP to 2.5% in 2027 via the relatively politically painless step of cutting international aid spending, but finding further money will prove harder. France’s government collapsed in January of this year after proposing cuts to social security after ratings agencies had started to downgrade French debt.

The German government is attempting to repeat its 2022 manoeuvre of suspending its notorious constitutional ‘debt brake’, which limits the federal government’s structural deficit to 0.35% of GDP, to facilitate greater defence and infrastructure spending. This is running into difficulties, as the Greens are threatening to block the move in an attempt to extort more money for environmental causes. A higher spending, higher borrowing German economy might well be more dynamic long-term, but this is already having implications for the borrowing costs. Yields on German bonds have spiked following their biggest sell-off since 1990.

This has implications for other European countries.

German bonds were historically considered some of the safest sovereign debt in Europe, to the point where it carried a negative yield at times. This means that investors were paying to lend the German government money (while counterintuitive, this can make sense). If bond investors can now expect a 3% yield on German bonds, they will expect higher returns from riskier countries. This has already started to push up French and Italian borrowing costs.

With the political mood against spending cuts, are tax rises on the agenda? With the tax burden in many European governments already at their highest in decades, large-scale tax rises will prove difficult. That said, some European governments are exploring squeezing their wealthiest citizens further.

The only realistic path forward is likely to involve a heavy debt component.

As our French and German examples show us, markets aren’t wild about individual European nations borrowing ever more money, the EU’s broad shoulders may yet play a useful role.

There has already been talk of the EU loosening some of its rules on government debt and shuffling money between various bank accounts, but the most controversial proposal is a plan to borrow up €150 billion for defence.

The EU has issued debt for a while, but has scaled this into the hundreds of billions in recent years. We saw this in the €650 billion Covid-era Recovery and Resilience Fund, as well as its Ukraine Facility.

Despite relatively high debt-to-GDP ratios in many individual European states, EU bonds have so far proved reasonably popular with investors. Some EU bond issues have been as much as 11x oversubscribed. These bonds benefit from the backing of multiple governments and the support of strong EU institutions and governance frameworks. Every ratings agency except S&P gives the EU an AAA rating.

Euro bonds also provide investors some helpful diversification, giving them a high-quality euro-denominated asset, without tying them to a single country’s risk profile. They also offer the “safe haven” status of the German Bund (issued by the EU’s only major AAA-rated economy), without its (until pretty recently…) measly yields.

If this is such a no-brainer, why hasn’t the EU already done it?

There are three blockers - the first of which can probably be (sort of) ignored, while the latter two are solvable with concerted political will.

Firstly, the macro backdrop for bonds is less favourable than it was in 2020-2021. Interest rates are higher, there’s been a flood of big government debt issues in the past year, meaning that borrowing costs are going up. That said, there’s early evidence of institutional investors expressing interest - it’s just likely to come at a steeper price.

Secondly, we should expect a blazing row about who does and doesn’t want to participate. In short, the EU is split into factions between the enthusiasts (e.g. countries with borders near Russia), the sceptics who would rather existing money was moved around in the first instance (e.g. The Netherlands, Sweden), and the undecideds (e.g. Denmark, Germany, Austria).

As a result, we may end up seeing a vehicle that operates outside the standard EU treaty framework, potentially via an intergovernmental agreement between participating countries. This would be unchartered territory, but might have the advantage of allowing the UK to participate. The more countries that join, the better. If the coalition of the willing ends up being very small, “I can’t believe it’s not French debt” is unlikely to set investors’ hearts racing.

Thirdly, we should expect a tussle over how the money is spent. France will lobby hard for the money only to be allocated to European defence contractors (I wonder why…), while other nations will want more flexibility.

New money, legacy commitments

Once the spending taps are turned on, should we expect to see European rearmament accelerate drastically?

Up to a point. Unfortunately, defence doesn’t start from a blank slate.

For example, France did hike defence spending last year, with the country preparing to spend 40% more between 2024 and 2030 than it did between 2019 and 2025. While this sounds like a lot, it goes less far than you’d think. This is for two reasons.

Firstly, the more sophisticated your military, the more existing commitments need funding. In France’s case, 13% of the new budget will immediately be consumed by the nuclear deterrent, where carrier platforms, missiles, and submarines all need upgrading. It’s not an altogether different picture in the UK, where the government’s decision to prioritise the punctual upgrade of the nuclear deterrent will eat up a meaningful proportion of the money allocated to the Ministry of Defence’s equipment plan until 2033.

Secondly, during the post-Cold War lean years of European defence spending, governments routinely deferred modernisation programmes and stripped out resilience, creating significant legacy debts. This means new money often has to be spent fixing existing equipment or restoring empty stockpiles.

For stretches of Operation Barkhane, France’s counter-insurgency operation in Sahel, less than half of the country’s military helicopter fleet was deployable, while Tiger attack helicopter availability dropped as low as 25%. Meanwhile, since 2011, six out of the country’s 20 ammunition depots have been closed, leaving the armed forces with the stocks to last a few weeks in a combat against a peer nation.

France is at the better end. In 2015, equipment shortages forced German soldiers to use broomsticks painted black instead of guns during a NATO exercise.

Could we all just be more like Poland? Many defence enthusiasts I know look at the country enviously, as it prepares to increase defence spending to 4.7% of GDP this year. Poland has the largest NATO land army in Europe and has won plaudits for its efficient off-the-shelf procurement of South Korean tanks, as it transitions away from old Soviet equipment.

There’s a lot to commend about Poland’s sense of urgency, but the picture’s a little bit more complicated than it appears on the surface. Much of its budget goes unspent, there’s a shortage of experienced personnel (thanks to a combination of recruitment challenges and politically-motivated purges by the previous government), procurement is more chaotic than the headline orders make it appear, and there’s a shortage of specialists who can help implement new systems. In short, building an effective military and accompanying culture takes time. More money helps, but it can’t shortcut everything.

Subscribe now

A stretched industrial base

But let’s say we find the money to resolve all the legacy dents and then acquire some new capabilities, is that it?

Not quite.

You then need to produce and buy the equipment. Most defence contractors don’t have large stockpiles of spare munitions sitting around, especially the kind that would be useful for a confrontation with a peer adversary. Margins in defence are slim, as governments tend to cap profit levels. This means manufacturers are incentivised to run just-in-time operations optimised to the needs of their biggest customers - rather than maintain stockpiles or spare capacity.

This is even the case in a country like the US, where defence spending has remained reasonably high after the Cold War. Since the turn of the millennium, the US has been overwhelmingly focused on counterinsurgency operations and asymmetric warfare.

For obvious reasons, these operations tend to rely more on light, mobile equipment and precision munitions. Considering the Taliban and al-Qaeda were unlikely to ever outgun their American opponents, there was also little need to maintain much idle capacity. As a result, the US can reportedly only produce 1/9th of the shells a month that it could in 1995. A recent Taiwan wargaming exercise found that the US would exhaust its entire global inventory of long-range anti-ship missiles within the first week of the conflict.

While President Trump is insisting that European governments are obliged to buy more American arms, US manufacturers are already struggling to meet existing orders from foreign customers. This means there is little option but for European countries to restore domestic arms production.

Unfortunately, this is far from easy.

France again gives us a helpful illustration. A once globally significant defence industry has been allowed to atrophy due to lack of domestic demand and cheaper competition from countries like South Korea. While sales of the Rafale fighter keep French industry relevant, there is little domestic production of infantry equipment. The only notable exceptions are lightly armoured vehicles like the Jaguar and Griffon, which are designed for the kinds of asymmetric conflicts France has been fighting in west Africa.

The Leclerc, France’s main battle tank, is no longer produced at all, after it proved an export flop. The Main Ground Combat System - a Franco-German project to replace the two countries’ battle tanks - looks unlikely to be operational before the mid-to-late 2030s.

To compensate for these weaknesses in the country’s defence-industrial base, the French government is exploring whether car manufacturers can turn over some of their factories to the production of kamikaze drones.

Rheinmetall, Germany’s biggest arms producer, has won plaudits for its energetic programme of expansion - building or acquiring facilities across Europe. This capacity will be very useful … eventually. Building and equipping a new factory for heavy industrial production is a slow process, especially if your main customer moves slowly. Even in the face of the Ukraine war, Germany was slow to place new orders. The German armed forces took until July 2024 to order a batch of new Leopard 2 tanks, which will mean delivery in 2030 at the earliest.

Mass and infrastructure

Once the new equipment arrives, there’s the additional task of operating and transporting it. The two crucial components for this are mass and infrastructure. Europe lacks both.

Mass is critical - not just for fighting, but because of defence’s growing complexity. In a world of increasingly advanced weapons systems, aircraft, and vehicles, a long chain of people are needed to manage logistics and technical support.

The proportion of combat soldiers versus support, sometimes called the “tooth-to-tail ratio”, has grown from about 1:4.3 during World War II to closer to 1:10 today. This ratio was one (although not the most decisive) factor in the rapid collapse of Afghan security forces during the 2021 US withdrawal.

The Afghan military depended heavily on US contractors for maintenance, repair, and logistics. Once the US supply chains were dismantled, it was impossible to source spare parts. This, for example, meant that the Afghan Air Force was essentially inoperable. European governments risk being in exactly the same position with many critical capabilities; this is a much more concrete concern than the US flipping a hypothetical software kill switch.

There was a long-standing belief (or rather, hope), that new technology would reduce the requirement for mass. However, the Ukraine conflict has clearly demonstrated that technological advances in conflicts between peer nations are usually temporary. While Ukraine may have surprised Russia in the opening phase of the war with its use of drones, the Russian military quickly learnt to respond in kind. Mass quickly asserted its importance.

This is why a recent UK House of Lords report on UK military readiness concluded that “the evidence we heard points to the current size of the British Army being inadequate … we are concerned that the Army cannot, as currently constituted, make the expected troop contribution to NATO”. Meanwhile, Michael Shurkin, a long-standing analyst of the French military, argued in a report that despite the country’s military “being indisputably the most capable in Western Europe”, it “lacked the depth and the mass to do anything on a large scale for any length of time before it simply ran out of stuff”.

Once they have the equipment and the support staff in place, militaries then need to be able to move everything around. Crucially, you need good railways, roads, and bridges.

Germany, traditionally Europe’s key hub for military logistics, has underinvested in infrastructure, meaning it increasingly struggles to handle heavy infrastructure. There’s also a shortage of railway wagons capable of moving military equipment.

Once equipment gets through Germany, it runs into another problem. Lithuania, Latvia, and Estonia’s railway tracks were built with a wider Russian gauge, leaving the Baltic nation better connected to their prime adversary than their allies.

RailBaltica, the plan to connect Baltic capitals with Warsaw via European standard-gauge track won’t be completed before 2030 at the earliest.

So, is all hope lost?

After this litany of woe, it might be natural to counsel despair in the face of overwhelming odds.

While the kind of Ukrainian victory people envisaged in the summer of 2023 is clearly off the table now, the Ukrainians aren’t up against an invincible war machine. As former Ukrainian defence minister Andriy Zagorodnyuk put it in a recent article:

At the moment, Russia’s troops are exhausted, its defence production is struggling to keep pace with battlefield losses, and its unmanned weapon systems are, despite great effort on Moscow’s part, struggling to prevail over those of Ukraine. Unexploded drones and missiles recovered from the battlefield are mostly of very recent manufacture, indicating that Russia has depleted much of its stockpile and is deploying new weapons as soon as they become available. Meanwhile, even with Russia’s large population (and soldiers from North Korea), recruitment levels are failing to meet demand, and the high casualty rate leaves little time for adequate training of new troops.

Additional European support could make a difference in strengthening the Ukrainians’ hand, so they can secure a better outcome.

Beyond the immediate conflict, there is a bigger political argument to be won. As we established above, good defence requires long-term investment and planning. That means spending will need to be maintained well above 2.5% of GDP for years after any cease fire. The temptation to raid defence budgets to buy-off increasingly weary electorates, who might otherwise be tempted to vote for populists, needs to be resisted at all costs.

These battles are winnable. At the end of the day, Europe’s most immediate adversary is a country with a commodity-dependent economy smaller than that of Italy, a declining population, increasing dependence on China, and diminishing influence in much of the world. This doesn’t have to be David versus Goliath (although famously, David came out pretty well in that match up). But if European countries want to defend the values they so often champion, they’ll have to pay a price.

Subscribe now

Disclaimer: These are my views and my views alone. They aren’t those of my employer, my dentist, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Image credit: Leclerc battle tank, Daniel Steger (Lausanne, Switzerland), CC BY-SA 2.5 <https://creativecommons.org/licenses/by-sa/2.5>, via Wikimedia Commons

Introduction

If you spend enough time around enough founders and investors, you’ll quickly discover that one of the most popular explanations for Europe’s relative lack of big tech companies is ‘culture’.

Tom Blomfield, the co-founder of Monzo, articulates this view well:

“In the UK, the ideas of taking risk and of brazen, commercial ambition are seen as negatives. The American dream is the belief that anyone can be successful if they are smart enough and work hard enough. Whether or not it is the reality for most Americans, Silicon Valley thrives on this optimism”.

I’ve heard versions of this expressed by countless other successful people, but I’ve always found “something something American dream”, “in the US people build”, “there’s an appetite for risk”, “being wealthy is seen as aspirational” and the many, many variants thereof to be unsatisfying explanations.

In fact, 99% of the time, cultural explanations of economic divergence, especially when it comes to big company creation, just read like people projecting their personal ideological hang-ups onto either a European or American Other.

To be clear, I’m not trying to argue that culture doesn’t vary between societies. That would be silly. If we gathered a group of people from Amiens, Atlanta, Armagh, and Amman in a room - their outlooks on the world would likely diverge significantly.

But at the same time, this has produced an industry of takes that are practically … useless. My contention is that cultural theorising is a poor way of explaining Europe and the US’s divergent economic performance, ability to scale companies, and a lot of other things. It encourages defeatism and a fuzzy view of policy. This week, Chalmermagne goes Marxist and makes the case for remembering on the economic base.

Subscribe now

Explanation or ideological bias?

With the greatest respect, neither founders nor VCs tend to be gifted historians or macroeconomists. Nor do they need to be. This is why their analysis of economics often feels like vibes-based extrapolation from the traits of people they admire.

By their nature, the most successful founders are highly unrepresentative on just about every metric. If someone has built a multi-billion dollar company, they are freakishly talented and driven, intensely lucky, or a criminal. Or a combination of the above. Whichever way you cut it, they tell you next to nothing about the median person in their society. Thanks to the strength of the US tech ecosystem, Americans are likely over-represented in this sample.

These cultural explanations are then commonly repeated as if they’re self-evidently true.

For example, a common contrast you’ll hear is between American respect for success and British ‘tall poppy syndrome’, where us Brits have to cut down anyone newly wealthy we know due to a vague theory about the class system.

This conveniently ignores the widespread popularity in the US for greater taxation of the wealthy. Close to two-thirds of Americans have expressed support for a wealth tax on the super-rich. A similar proportion believe the gap between rich and poor is too wide. These aren’t that different to the statistics we see in the UK - with a pro-wealth tax organisation (a group incentivised to find the highest possible number) finding a similar proportion in favour of a tax to our US survey.

But even if the UK number was significantly higher, would this have really have significant explanatory power? Are the people who build generational companies actually dissuaded by a loose sense of their neighbours’ jealousy?

We can apply the same test to a bunch of other cultural markers.

Let’s take individualism.

Americans are indeed more likely to believe their fate is in their own hands … by a whole two more percentage points than the UK and seven than France. Maybe those two points make all the difference, but colour me sceptical. The US beats Poland by 23 points on this measure, but Poland has also been one of the continent’s great economic success stories. Self-reported individualism might be an example of a cultural difference, but does it predict anything useful?

Do Americans have more of a work ethic than Europeans?

Slightly, but not as markedly as you’d expect based on popular perception. And significantly less than the anti-individualist Poles we were discussing above.

Three economic factors in a trench coat

Now, of course there are important differences between America and European countries. But to what extent are ‘cultural differences’ really best explained ‘culture’?

Let’s take employment rights as an example.

It’s true that the US has much weaker employment protections than essentially any European country. You can assert with a reasonable degree of confidence that strong employment protections come with a trade-off around business dynamism.

If it’s time-consuming and expensive to get rid of people who underperform or who you no longer need, you’re likely to be more cautious about hiring. Similarly, if your company has to pivot to a new business model or technology, stringent employment laws make it hard to adjust your team accordingly.

The cultural explanation would be a version of: people in European nations place a premium on stable employment and they are prepared to sacrifice a degree of economic dynamism to preserve this.

Now, there’s probably some truth in this. If you give people a lot of employment protections or subsidies, they will be reluctant to hand them back in exchange for a hypothetical increase in GDP growth. This isn’t a uniquely European characteristic - just look at the unreformed and increasingly unaffordable US entitlement system. The frontier spirit apparently doesn’t extend to social security or Medicare.

This is because path dependency, rather than culture, is likely a better explanation of the difference.

You see this really clearly when you dive into individual examples. Denmark is known for its relatively unusual system of ‘flexicurity’, where employers are able to hire and fire easily, while the unemployed are able to benefit from a generous safety net. This means churn is reasonably high in the Danish private sector.

But Denmark otherwise has a more typical ‘Scandinavian’ welfare system. This suggests ‘culture’ is a weak explanation, so how did it end up in this unusual position?

In short, after some particularly ugly strikes and an ensuing lockout in 1899 essentially drained both the unions and employers of resources, the two sides agreed to a compromise where employers retained flexibility in hiring/firing decisions while unions gained influence over wages and working conditions through collective bargaining. This framework was established outside the government. Meanwhile other European countries tended to be left with a more binary choice between strict employment protections or limited welfare provision.

The US, in turn, benefitted from … a combination of the New Deal and not being economically decimated by World War II. As European nations rebuilt their economies and constitutions from scratch, scarred by the memory of the Great Depression, they were sympathetic towards social-democratic policies and strong unions. The US felt no need to change course and give the workforce a greater role in economic governance.

Cultural factors also struggle with time - there have been moments when US and European performance on any number of economics have aligned, and times when they’ve diverged. During the mid-to-late 90s, for example, both regions saw remarkably similar patterns of strong economic growth, low inflation, falling unemployment, and strong equity markets. Can culture really have changed so much in the last couple of decades?

If we go back to the Tom Blomfield quote at the beginning, is risk tolerance really best explained by culture?

His piece opened with him questioning why talented US graduates seemed more inclined to work in start-ups, while their British peers tended to take jobs at McKinsey and Goldman Sachs. The vibes in Oxford or Cambridge are almost certainly different to those in the Bay Area. But it’s also true that the US has the world’s most mature technology ecosystem, with a large number of high-profile success stories. Smash hit UK start-ups and repeat founders are rarer, while the UK is one of the leading markets for financial and professional services.

But why not use ‘cultural factors’ as a short-hand for the effects of these historic economic forces?

In short, because it isn’t helpful. These differences were the product of policy decisions, not cultural vibes. That means that they can be changed via politics if necessary. There is nothing ineffable about European-ess that leads to a specific regulatory regime for employment rights, or basically anything else. For the vast majority of economic issues, ‘culture’ leads us into an analytical cul-de-sac.

Uninventing economics

Even if we shift a few policies, with reforms to procurement here, or changes to employment law there, Europe is still confronted by the glaring challenge of … reality.

In the discussions about companies not scaling domestically or fleeing overseas, investors and founders often forget good old-fashioned economics.

Comparative advantage, formalised in the early nineteenth century, accepts that trade patterns are generally determined by relative rather than absolute productivity differences. Even if one party is better at everything (absolute advantage), both parties benefit when each specialises in what they're comparatively best at.

It will often just make more sense for entrepreneurs to scale companies in a country with a bigger market, looser employment laws, lower energy costs, and the vast majority of the world’s best academic institutions. As a result, people are more likely to want to raise growth rounds from established US funds that can help them plant roots on the other side of the Atlantic.

Now, Europe could address some of the underlying causes of the US’ comparative advantage in scaling.

It could, for example, increase the size of its market by tearing down its internal trade barriers, which according to the IMF, are the equivalent of 45% for manufacturing and 110% for services. There’s currently the EU Inc. effort underway to campaign for a Europe-wide entity for start-ups.

Additionally, EU member states could push to weaken overbearing tech regulation, whether it’s GDPR or the EU AI Act.

European countries could fire up nuclear power stations to bring down persistently high energy prices, unless they fancy another stint of Russian natural gas dependence.

But to do so, it needs to decide why it wants to and develop an underlying theory of how to incentivise and foster big company creation. There are perfectly reasonable theories about why both of these are good, but I’m yet to hear many European governments articulate them. Instead, they’ve opted to pump more and more money into a pale imitation of the US venture ecosystem, so it can sit in more undifferentiated SaaS companies with a limited footprint.

Culture or legacy?

Having said all of the above, I regrettably don’t believe that if we just pull a few economic levers, everything will magically be fine.

There are certain biases that need to shift that I will grudgingly accept are cultural. I distinguish these from the policy and economic changes we covered above, because there aren’t simple policy levers that can be pulled. They’re ingrained habits that aren’t downstream of political decisions. These will likely only be resolved by a change in mindset.

A major bias is the mercantilist way many European governments, academic institutions and start-ups approach intellectual property. They routinely mistake IP hoarding for value capture.

For example, in the UK, the House of Lords Communications and Digital Committee recently published a report, which warned that “we have consistently struggled in enabling our brightest technology startups to transform into significant domestic businesses and potential British world-class companies … the consequences of this failure are significant”. Baroness Stowell, who chaired the committee, wrote in an opinion piece that:

There can be no doubt that British tech unicorns galloping overseas as they reach maturity is damaging to UK PLC. Decreased global competitiveness, weaker economic prospects and a ‘brain drain’ of talented individuals are all likely consequences of failing to support our companies to scale at home.

This is a serious problem, especially for a Government that has declared economic growth to be its number one priority. We can and must do better than act as an ‘incubator economy’ for countries with greater ambition.

The ‘incubator economy’ framing isn’t just theoretically unsound, it also overlooks how the economic benefits of innovation diffuse. Historically, only a miniscule proportion (~2%) of the gains created by technical advances are captured by their producers, with the rest going to consumers. After all, the creators of the printing press were not the main beneficiaries of its invention. In fact, early adoption of a German invention in Holland led to local innovations that were in turn exported back to Germany.

This is because (with the exception of a few industries) most innovation is expensive to produce initially, but comparatively inexpensive to reproduce, adapt, and distribute. This is particularly obvious in a field like AI. If you have the budget and the team, it’s often surprisingly easy to reverse engineer other people’s breakthroughs. We’ve seen this with the run of reasoning models following OpenAI’s o1, along with rapid replications of Google DeepMind’s AlphaFold 3.

This isn’t unique to the UK. A different spin on the IP problem emerges in Wolfgang Münchau’s Kaput, which documents the souring of the German economic miracle. In Münchau’s telling, the historic success of Germany’s industrial heritage and export-led growth model bred intense complacency. The German car industry, for example, did everything it could to resist the shift towards electric vehicles and the growing importance of AI. German incumbents prided themselves on their international leadership in … the number of patents and responded to fresh competition via a combination of cheating and calling for tariffs.

You even see a version of this in some European start-ups. A number of the continent’s most high profile anti-success stories were the result of companies concentrating too much effort on their IP and ignoring such prosaic questions as customers and product. Perhaps the peak example of this is Graphcore, the AI chip company, which combined elegant hardware with janky homebrew software. This is a mistake that NVIDIA would never have made.

This bias in turn leads governments to pursue value-destroying policies, designed to keep innovation domestic - whether it’s exit taxes, M&A restrictions, or ever-greater subsidy of the venture market. The sense of relative decline and faded past glory is a powerful drug.

Closing thoughts

So where does this leave us?

There are broad strokes variations in cultural bias that explain underperformance in certain sectors. People with those attitudes should get swept away with time. But ultimately, path dependency is real. What seems like an ineffable vibe is often better explained by an economic crisis, bad policy decision, or 19th century collective bargaining agreement you’ve long forgotten about.

Next time you want to comment on a trend: ask yourself, is this best explained by policy, institutions, and history, or by collective bias?

Maybe it’s a bit of both.

But as a moderate, temperate Englishman, I would say that, wouldn’t I?

Disclaimer: These are my views and my views alone. They aren’t those of my employer, people I run past in the street, or anyone else. I’m not an expert, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Subscribe now

The late Polish philosopher Leszek Kołakowski was mainly known for his detailed work on Marxism and theology. But in 1986, he wrote a short thought experiment for literary magazine Salmagundi. Entitled “The Emperor Kennedy Legend”, it envisioned a world, thousands of years in the future, where most traces of humanity’s past had been wiped out by a Great Calamity. Only a handful of books and academic journals from the previous era survived the floods and explosions.

Assembling scraps of information about the life of John F Kennedy, various historians and anthropologists had concluded that the Emperor Kennedy was a likely mythical figure that primitive people in the Before Times worshipped. Kołakowski describes how approximations of Freudians, Marxists, and structuralists in the post-calamity world projected their own theoretical interpretations onto the three or four pieces of evidence about the Kennedy Legend. Maybe Emperor Kennedy’s life was a reflection of class struggle, binary male-female oppositions, or the eternal fear of castration?

At no point in the debate, do any of the participants stop to question their assumptions, whether or not their shared understanding of the facts is true, or if there may be some evidence that they’re missing.

Policy brain

When I’ve written critically about current government policy, friends have reasonably asked “so, what would you do instead?”. On some questions, my answer is “well… nothing”.

If you write about policy for a living, you are predisposed, unless you are one of the world’s six remaining libertarians, to believe that most problems have policy solutions. There is a lever, which people in power can pull, to ameliorate problems or drive progress.

My contention is that this outlook can produce a bias towards action, any action - irrespective of whether it’s actually helpful.

A good example is the debate around why so few British start-ups scale domestically. In my view, scaling challenges reflect a mixture of economic logic, ecosystem immaturity, and the lack of repeat founders or founders turned investors. This will probably improve by itself, to some degree, over time, and much of what we do in the intervening period is expensive displacement activity.

If you look at this challenge through the lens of what I call Policy Brain, the prospect of waiting becomes unbearable. You’ll desperately scrabble around to find some levers to pull, even if few good ones exist. This will result in commissions, taskforces, and new pots of money being pulled together. A mafia of vested interests will coagulate, using the scaling challenge as an excuse to lobby for money or advantageous policy change.

It’s the economy, stupid

Policy Brain, however, isn’t just the impulse or blindspot of some idealistic nerds. It’s a reflection of economics.

Ultimately, research costs money and someone needs to provide it. That may be a donor or a sponsor for a think tank, or a fixed budget for an in-house policy team.

The people spending this money expect it, directly or indirectly, to have some influence on the world. This does not incentivise the publication of research rejecting interventions or proposing masterly inactivity.

Many researchers will naturally assume that this is either not what their benefactors or their boss are paying for. “We took a long hard look at X and while we agree it’s bad, we’re not convinced we can do much about it” or “we’ve had pet theory Y for ages, we tested it and by every conceivable metric it sucked” are both hard sells for different reasons. This is one driver of the huge volumes of low value policy work that Anastasia and I wrote about last year.

But it doesn’t have to be this way.

We’ve seen this movie before

The closest parallel to the problem of policy filler comes from the world of scientific publishing. There has been a long-standing push to encourage the publication of more null results - research where the findings don’t support the hypothesis being tested. One 2022 survey suggested that 75% of scientists were willing to publish null results, but only 12.5% had been able to do so.

In short, this is because peer-reviewed journals see them as bad box office.

In a world where academic researchers are cautioned to publish or perish, this creates some obvious perverse incentives.

This has led to widespread use of practices such as p-hacking, where researchers misleadingly produce statistically significant findings. This is actually surprisingly easy to do. One form of p-hacking is the misuse of subgroup analysis - where you examine the effects of an intervention or treatment on specific subsets of study participants (like age groups, genders, or other demographic categories).

If you analyse enough subgroups, you are likely to find some statistically significant results by sheer chance. For example, if you test 20 different subgroups at a 0.05 significance level, you have a 64% chance of finding at least one "significant" result, even if there's no real effect. Minus out the other 19 groups, and voilà, your statistical noise has become a valuable new contribution to the literature.

This practice is a staple of the bad research The Studies Show, one of my favourite podcasts, digs into on a regular basis.

This kind of bad practice helped fuel the ‘replication crisis’ - the dawning realisation that scientists are simply unable to reproduce reams of eye-catching findings. Power posing, much of the Stanford Prison Experiment, and the brain-enhancing effects of listening to Mozart all fall into this category. This phenomenon started in psychology in around 2011, but has spread much further.

This has led to a push for the greater publication of null results, along with the use of pre-registration, where researchers publicly state their hypothesis and how they plan to test it, before conducting the work. This makes it harder (although not impossible) to magic a hypothesis into existence retrospectively through statistical sleight of hand. This does of course comes with trade-offs, such as the potential loss of the serendipitous discoveries that come with open exploration.

There’s also growing interest in the idea of Registered Reports. In this system, researchers get feedback on their methodology before conducting research. Provided that it is then executed faithfully, the journal will publish the resulting study, even in the event of a null result.

This isn’t just good for integrity, it advances the field. If you work in any discipline, you might be interested to know if someone smart has constructed a hypothesis, only for the evidence not to support it. After all, this null result might inform how you approach the same question.

Losing this data to a tidal wave of filler and spurious correlations impoverishes everyone.

Politics vs policy

One disadvantage that policy suffers from that (most) science (usually) doesn’t is the impact of politics. Legislators want to legislate. The bias is almost always to do something, or at least to be seen to be doing something. Earlier in my career, I was given the advice that if you wanted to stop a politician from passing a law, it’s not enough to prove it’s a bad idea. Instead, you have to find something else for them to do.

This impulse has produced no end of questionable legislation. A recent example is “Martyn’s Law”. The Terrorism (Protection of Premises) Bill 2024 is legislation currently under discussion in the UK, which was drafted in response to the 2017 Manchester Arena Bombing.

The inquiry following Manchester found that the attack could have been prevented. The security services were overstretched in the run-up, and there were a number of critical failings on the night. These included police officers disappearing for twice their allotted break time, a security guard failing to escalate a report of a suspicious person, and another security guard being too afraid to approach the suspect out of fear of appearing racist.

This is all deeply regrettable. Unfortunately, the security services can make mistakes. Police officers and security guards can make bad spontaneous decisions. Organisations should urgently study what went wrong and draw lessons from it. But not everything bad that happens to the public is a public policy question. Laws can’t drive the rate of bad decisions by individuals down to zero. Nevertheless, the government decided to legislate.

The resulting bill imposes a tiered set of requirements on venues, starting at those capable of holding over 200 people, to introduce measures to reduce the likelihood of terror attacks and implement response plans. At the cheap end, this will mean more form-filling and training. At the expensive, new infrastructure. The burden would disproportionately fall on smaller, community venues.

The government’s own assessment of the policy’s Net Present Social Value (its metric assessing the costs and benefits of a policy) suggested that it would contribute a cool negative £1.8 billion. To be clear, this measure takes into account the potential lives saved.

Martyn’s Law isn’t the only example of ‘something must be done’ policy-making. You could count the Online Safety Act, the independent football regulator, regulation on window size in newbuilds, proposals to blunt kitchen knives, and almost every policy recommendation to fight misinformation ever. We’re really good at this.

Closing thoughts

Of course, people working in policy can’t stop politics. But they have a responsibility not to fuel the flames with yet more filler. A proposed extra form or assessment here, an additional compliance check there, or a new government office for counting badgers can all have unexpected and costly downstream consequences.

Null results can be useful for the policy practitioner as well as the scientist. If a team you respect has conducted an economic analysis or consulted people with frontline experience, and concluded that there isn’t a short-term fix, the real problem lies elsewhere, or a fashionable intervention may be less effective than it appears - that’s useful.

Policy people should also have enough confidence in the value of their work to make this case to their employers or donors. Publish your ideas that didn’t work out on closer examination. Say when you think there’s no good answer. Otherwise, we may as well spend our time attempting to crack the Emperor Kennedy Legend.

Disclaimer: These are my views and my views alone. They aren’t those of my employer, traffic wardens in my local area, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Image credit: data dredging - https://www.tylervigen.com/

PS: if enough people send me some good null policy results that they haven’t been able to use, I’m up for publishing a collection. Hit reply or DM me.

One of the best parts about writing Chalmermagne for the past couple of months has been the thoughtful responses I’ve received to my weekend scrawlings. While it’s always nice to get praise, I particularly appreciate the people who write in to explain why they think I’m wrong or where I’ve missed a trick. While I don’t always agree, I also realise that I can’t hope to see every possible angle of every issue, so it makes me sharper.

Every now and again, as some filler content to build reader community, I’m going to share a summary of the most interesting critiques, along with a few reflections.

Keep it coming!

Subscribe now

Defensiveness about inoffensiveness

Seemingly people couldn’t get enough of my policy collab with Anastasia. I’ve heard stories of it being brought up at the start of roundtables … only for them to proceed along exactly the same pointless lines. As we warned in the piece, ‘winning mindshare’ isn’t the same thing as achieving change…

Chalmermagne

Death by a thousand roundtables

For this piece, I’ve teamed up with Anastasia Bektimirova, a researcher focusing on technology and science policy. Anastasia writes in a personal capacity, so nothing in this piece should be seen as representing the views of her current or past employers. You can keep up with her work on X at…

Read more

2 years ago · 88 likes · 3 comments · Alex Chalmers and Anastasia Bektimirova

But Martin Koder, currently AI Governance Leader at financial messaging service Swift provided a thoughtful defence of one of our maligned archetypes:

All rings true apart from underselling the second archetype “The government is already doing it … and we agree”. This is actually an incredibly effective way of getting policy changed because you can reframe your ask as a better way to help policymakers supposedly deliver the objectives they already have. Such papers always begin with the same line "We welcome the [policymaking body]'s proposal for X. To help it deliver on its objectives of [motherhood and apple pie] we recommend some technical recalibrations to the [key threshold that impacts our bottom line] to ensure alignment with the [current meta policy fad of the most senior politicians in that jurisdiction] and consistency with [inevitable overlapping extant policy controlled by a different departmental head]. We present some new data demonstrating how bringing the threshold into line with [some other arbitrary classification already coded into law which happens to suit] could generate up to [gerrymandered but superficially defensible number] extra jobs 'n growth. I got swathes of legislation changed like that.

There’s probably a lot of truth to this. Framing your ask through the priorities of the people you’re trying to influence is obviously sensible. But the key here is having an ask.

Too much policy work unfortunately plays back stakeholders’ priorities and then forgets the next bit. This was the genre Anastasia and I really had in our crosshairs with this one.

I’m loving angels instead

Life hack: if you ever feel lonely, publicly call for a few tax breaks to be removed. You won’t be short of correspondence for weeks. My critique of the Enterprise Investment Scheme and Venture Capital Trusts - vehicles used by wealthy individuals in the UK to invest in high-risk companies while reducing their tax bill - provoked an avalanche of feedback.

Chalmermagne

A parody of venture

Introduction…

Read more

a year ago · 17 likes · 5 comments · Alex Chalmers

Helpfully for the purposes of a summary, most of the criticism tended to focus on the same handful of points.

For example, Fred Soneya, Co-Founder and General Partner at Haatch covered two of the most common arguments: that I overlooked good funds and wasn’t harsh enough on bad practice in traditional UK venture.

I think some of your writing is spot on, and some is way off. That's coming from someone who manages a mix of capital, including SEIS and EIS.

I started to draft a response in full, including the areas I agree with you on (like fees—we charge no fees to portfolio companies) and areas I disagree on ("the incentive of an EIS fund manager is to deploy as much capital as possible and to protect the downside"—we delivered a 276x return, which was not an investment made with any view on how to protect the downside; we were all in personally), but I ran out of characters!

I also think it's easy to point the finger at the EIS Industry on points like fees, returns, incentives, and alignment. However, I can give you the same number of examples of GP/LP and PE Funds, which do precisely the same thing you're calling these managers out on.

There are also incentives that have been created (BBB ECF) that provide quasi-tax relief for investors to invest in venture funds, but are you saying that is okay versus EIS?

Meaning no offence to Fred, I didn’t find either of these arguments hugely persuasive.

Firstly, incentives aren’t the same thing as destiny. There can be good actors in a marketplace that incentivises bad behaviour, but they’re likely to be under-represented. We should commend the team at Haatch for not charging investee companies fees, but data in the piece showed that i) most EIS funds do charge fees ii) EIS funds are likelier to charge high fees than non-EIS funds.

In much the same way, a 276x investment (however impressive) in one EIS fund doesn’t negate lower returns among EIS funds in general.

To convince me that this argument is wrong, I would need to see evidence that I had either misunderstood the incentives or that EIS funds as a whole were much better actors than I realised. No one so far has made that case to me, but I’m open to it! Pointing to two or three better funds (as I did in the piece), however, doesn’t prove much either way.

Secondly, the attempt to turn my argument back on traditional UK VC overlooked that I’d made exactly this point in the piece!

Much of the bad behaviour I’ve described in this piece isn’t confined to EIS and VCT. There’s a longer piece to be written looking at how bad practice, founder unfriendly behaviour, and short-termism remains prevalent in the wider UK ecosystem.

If anyone thinks I’m too soft on the rest of the sector, they should check out the piece this Substack launched with. On day 1, I had a mailing list of … three people, so you are all forgiven for having missed it.

Chalmermagne

A bridge fund to nowhere?

Introduction…

Read more

2 years ago · 14 likes · 5 comments · Alex Chalmers

The most interesting critique came from Dom Hallas, the Executive Director of the Startup Coalition, which lobbies (very effectively) on behalf of the UK early-stage tech ecosystem. Professionally, I’ve been on the same side as Dom in many policy fights, so I’d have been surprised if he’d agreed with my desire to heavily curb a scheme many of his members have used.

Dom made several different arguments in an extended thread on X, which you should read in full, but two jumped out at me.

It misses the founders who are scrabbling to build and raise, not just the 1% that get transatlantic venture money at seed stage who will probably always get funded. It’s about building the widest funnel of risky early companies possible and then seeing which ones make the leap. A ‘startup Gini coefficient’ would show this ‘haves’ and ‘have nots’ situation for founders and investors. 11x didn’t need a £250k SEIS round. But the haves aren’t the only people who could ever build businesses or invest in them. It’s great there are super successful founders and companies doing well - but try doing with no track record in the north of England and see how many Tier 1 term sheets you get.

I think the reference to transatlantic venture money is a bit of a misnomer, given that UK funds have raised tens of billions of pounds by themselves. The framing here risks implying that if you’re based in the north and Sequoia doesn’t come knocking, SEIS is your only other option. This might have been true of the 2004 UK ecosystem, but it rings less true in 2024.

At the risk of committing real venture heresy, I’m not sure we do need “the widest funnel of risky early companies possible”. I think we need a wide one, without artificial barriers, but funnel-widening isn’t free. The contribution of success stories versus the cost of the tax breaks and the money sunk into failures is something that we can measure. But, as I note in the piece, successive governments have shied away from robust economic evaluation in favour of polling the recipient of subsidies.

The other argument that jumped out was that funds are different. As Dom put it:

More businesses are taking ‘venture’ money and more funds have been created. They aren’t all doing the same thing and some of them don’t look much like the prototypical venture fund at all (though neither does Tiger Global or SoftBank). Ultimately this is probably a good thing - I think we’d all agree the financing for B2B SaaS businesses might not suit hardware driven climatetechs. But it clearly has quirks. Not all funding is the same (just like getting money from Sequoia vs getting it from a successful mid tier EU venture fund isn’t).

Tal Feingold, who coincidentally used to be with the Startup Coalition made an adjacent point:

The question I have is why is it bad to have a second tier of venture funds that aim for a lower return? Are you suggesting that the only companies worth funding are those that fit the trad venture model of hypergrowth?

While there is indeed more diversity in venture funding models, this doesn't excuse or explain why EIS funds specifically are underperforming. Many successful companies don't follow the “hypergrowth path”, but that doesn't mean that investors should accept poor returns. A climate tech start-up might generate big returns at a different pace to a B2B SaaS company, but to justify the high-risk and illiquidity of venture investing, the returns still need to appear at some point.

This is why ‘the second tier of funds’ doesn’t really wash with me: once you stop targeting outsized returns, the wheels start to come off the venture model. But if the product you’re selling isn’t “venture capital” and is instead “30% income tax relief”, then the calculation suddenly becomes different…

But these questions do point back to something I touched on in my first piece on this Substack: is the government too focused on VC-shaped solutions for early-stage businesses, even when they aren’t appropriate? Despite VC’s mindshare, it’s actually a pretty unusual way of financing a new business and isn’t the right vehicle for the vast majority of companies that are set up in the UK (or anywhere else).

Strikingly, all of the dissents focused on EIS. I haven’t received a single email, comment or reply defending the honour of VCTs.

Do you work at a VCT and feel maligned? Do you have a friend who works at a VCT who’s “alright when you get to know them”? Get in touch!

Defenceless?

Absolutely no one has written in to contradict my account of the poor state of UK defence. So either i) it’s as bad as I thought ii) there’s a secret plan they haven’t told us about but somehow a subset of Chalmermagne readers are in on it.

Chalmermagne

It’ll be alright on the night?

Sir Humphrey : Bernard, what is the purpose of our defence policy…

Read more

a year ago · 14 likes · 1 comment · Alex Chalmers

Based on my interactions with the Ministry of Defence to-date, I’d bet heavily on the former.

Exch-ch-ch-ch-changes

Last and, well, probably least the way it’s trending, is the fate of the London Stock Exchange.

Chalmermagne

London’s Stock Exchange is falling down

Introduction…

Read more

a year ago · 7 likes · Alex Chalmers

James Clark, from Molten Ventures a (London-listed) VC firm was first to step to the plate, pointing out some of the reasons for the decline:

Hmm. Interesting but you've missed a number of fairly substantial contributing points:

- Regulatory compliance lumped onto UK listed companies is substantially heavier than the US. The LSE will point out SarbOx compliance as a reason not to list in the US, but a whole range of ESG and DEI compliance points have been imposed on UK listings in the last decade. This makes life as a UK listed company less attractive when combined with...

- Historic shift by the funds management industry away from equity and into debt. Equities used to be 50% of holdings of UK pension funds managers, it's now something like 5-6%. This huge shift has dragged down prices for UK listed companies and less money chasing that listed companies will drive down valuations.

- the final point which you've sort of addressed is the tendency for pretty much everyone to conflate the entire UK capital markets with "LSE". I guess LSE being a local monopoly doesn't really help here. But the lack of clarity makes finding the sources of the problem hard while offering the exchange itself as a convenient whipping boy.

James framed his response as disagreement, but I don’t dispute any of this - these likely all are contributing factors to the LSE’s decline. My purpose in the piece was less the why, which has been analysed exhaustively, and more the so what.

Even if we make our peace with the LSE’s poor record on tech, James makes the important point that we can’t take much comfort in its performance when it comes to old-school industrials, its historic bread and butter:

The other key point here is that tech specifically is a pretty global market and that global market is centered on the US, with exchanges in the US consolidating the gains of global tech companies. Other industries consolidate in different markets, LSE being a focus for mining and resources. With this in mind, LSE losing BHP Billiton in 2022 was a far greater loss than ARM or any single tech company.

Rodolfo Rosini, co-founder of UK-US chip start-up Vaire Computing, thought I was too blaisé about the decline of the LSE, and stressed the importance of an equity culture for policymakers and the public:

When I moved to the UK 20+ years ago, I made a list of countries and one of the key requirements was to have his own stock exchange. I do not believe you can build a credible ecosystem without some form of public markets. Even if you decide to list elsewhere (eg Israel). It is important because having an equity culture understood by the public and policymakers is super important. The UK is going in the opposite direction.

I don’t want to put words in Rodolfo’s mouth, but I assume that by ‘equity culture’, he means that a strong exchange with high levels of retail participation should create a healthier policy environment, as well as making it easier for companies to raise capital.

I think this tallies with the point I was trying to make about Sweden at the end of the piece. Its domestic exchange is strong, especially for smaller companies, even though runaway successes like Spotify have listed elsewhere. Israel, which Rodolfo cites, is similar. I think this is probably the ideal end state for the LSE (rather than trying to win over Arm-style tech listings), but I accept that we’re politically and culturally some way off from this happening. In my eye-rolling at old school asset managers, I probably understated this.

When it comes to how we bridge that gap, we should probably engage with Zachary Spiro’s work on LSE reform. I referenced his work with Onward and the TBI in the piece and he responded with a thread of his own.

He argues that I’m too optimistic about the prospects of smaller UK companies on US exchanges and that sub-index companies will likely get lost:

Take analysts:

The average S&P 500 firm - 23 analysts

The average Russell 2000 - 6

Russell 2000 firms w/ $50m-500m market cap - just 4

What does this mean? It means that you can't just list as a sub-index UK company in the US and expect buckets of analysts/liquidity to follow you around. If you want US markets to pay attention to your IPO, you need a good reason for them to do so.

He points out that:

UK firms that list in the US *look different* to the ones that list in the UK.

They have more US revenue, and have probably already built relationships with US investors. [...] there will be some UK companies where listing in the US isn't the right thing - e.g. maybe their revenue is too EU-based.

My framing about the decline ‘not mattering’ was probably a touch flippant and Zach’s point is basically fair. A company that reaches IPO without significant connection to US investors will probably fall into that ‘well-suited to a Swedish-style exchange’ model. This is simply because few companies set to be global venture-scale winners usually reach IPO with no relationships with US investors.

Hopefully, the UK adopts the reforms that Zach’s paper outlines - alternatively, the companies could just go and list in Sweden. As the FT long-read on Sweden’s stock market notes, they’re not choosy about wooing small to medium sized foreign firms. UK asset managers take note.

Bonus segment: maybe the real regional subsidies are the friends we made along the way

My work off the stacks about UK tech policy didn’t attract a universally positive response. The piece contended that the government didn’t approach tech policy as tech, but instead treated it as a poorly-defined extension of economic policy.

Tom Forth, the CTO and Co-Founder of The Data City, took exception to my suggestion that this was primarily regionally-focused:

Don't get it at all. The claim is that "the UK has spent billions to build a machine for distributing regional subsidies" and yet the two examples of what we've actually built are in central London. Where is the regional subsidies? When do I get my regional subsidies in Leeds?

It’s worth reading his longer follow-up piece, which makes the argument that opponents of regional R&D subsidies “don’t think that putting something in London by default is regional development activity. They should.”

He notes that:

In data, tech, and now AI the UK government has consistently tilted the playing field in favour of London. The Open Data Institute, Tech City, Tech Nation, The Digital Catapult, The Centre for Data Ethics and Innovation, The British Library, The Alan Turing Institute, The AI Safety Institute, ARIA, and much more have been placed in the capital. The indirect government subsidies on offer in London make it ever more rational for British talent and ambitious companies elsewhere in the UK to relocate there.

I have a degree of sympathy with this, and in retrospect, I would probably take the word ‘regional’ out of my original sentence. This was lazy drafting and probably stems from staring too much at “regional venture capital funds” and other doomed government interventions of the 2000s for my first Substack piece.

I think the UK is a stupidly centralised country, which has negative economic and political spillovers that you can’t miss. I’d be up for dynamiting a significant proportion of the ‘London subsidies’ on Tom’s list and sowing salt where they stood, simply because I don’t think the orgs achieve much. But in my new devolved, smaller state Britain, am I up for a better geographical distribution of institutions? Sure.

Subscribe now

I can’t promise that masochistically reviewing all your criticism will be a very regular feature, but if you like it, hit reply and let me know. In the meantime, I probably have one more piece left in me before the end of the year.

See you next time!

Disclaimer: These views are my views or those of the people leaving comments. Unless they left them ironically or didn’t mean them, in which case, they should put more thought into how they present themselves in public forums. But more importantly, these aren’t the views of my employer or anyone who didn’t leave a comment. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Introduction

It’s been a bad few years for the London Stock Exchange (LSE). Over the past 5 years, the FTSE has climbed by a distinctly anaemic 15% , while the US S&P 500 has surged by over 90%. Just Eat Takeaway recently announced that it was delisting, the potential IPO of controversial online fashion retailer Shein led to the UK being dubbed the ‘market of last resort’, while the organisation that runs the exchange is widely seen as distracted and focused on other ventures.

There’s a sense of big opportunities being missed or lost. Deliveroo’s 2021 IPO was one of the worst in LSE history. London-listed Darktrace was snapped up by US private equity. Meanwhile, last weekend, London-based asset manager Nick Train said that Arm’s decision to sell to Softbank and delist “objectively was a mistake” and that it could now be one of the top 5 companies on the LSE.

I find the LSE discourse interesting, because I think it’s a good example of how policy discussions get mangled when a bunch of loosely connected issues are blurred together - either by vested interests or policymakers who, bluntly, don’t understand the mechanics. There are legitimate debates to be had around Britain’s poor recent record in large company creation, but the sound and fury over the LSE is largely a distraction.

This week, I run through why we should probably worry less about the movements of the FTSE and take a more realistic view of the prospects for big company creation in the UK.

People may just not know what an exchange is

One of the things that jumped out at me when reading historic coverage of the LSE is that some people may simply not know what listing a company on an exchange does and doesn’t entail.

To give a prominent example, the original Arm sale and decision to list in New York was pretty unpopular. One of the concerns raised was around sovereignty and national security - with a number of Conservative MPs leading the charge. After all, the company’s CPU core design sits in 90% of the world’s mobile phones. For example, Anthony Browne, then the Conservative MP for Cambridge South, said:

“Arm is a leading British technology company of national strategic importance and a major local employer. If it is floated on the stock market, it should do so in London rather than New York or elsewhere to ensure its interests and those of its investors are aligned with our national interest. Ownership matters – particularly of such strategically important companies.”

This is an odd argument for a couple of reasons. For a start, Softbank was only listing a minority stake in Arm. So its IPO venue would make no difference to who controlled the company.

More importantly - where a company's shares trade has limited bearing on who can own them or how the company operates. Countries like Vietnam or India have some restrictions about foreign ownership, but western capital markets are generally open. It’s usually easy for investors from anywhere to buy shares regardless of which exchange they’re on.

It’s not just ownership - where a company lists has minimal consequences for where a company employs people, where its tax revenue goes, strategic decision-making, or regulatory oversight.

If you want to upset some newspapers, you could even argue that by providing UK-based companies with access to deeper US capital markets that we’re advancing our national interests.

Atlas shrugged?

If we’re not worried about the impact of a declining LSE on national security, tax, or employment - should we be worried about its impact on financial services, given the sector’s importance to the UK economy?

In short, no.

The decline of the LSE isn’t great if you’re a tier 2/3 UK investment bank providing services to mid-cap companies or an asset manager with a predominantly UK mandate.

These may be noble professions, but they’re not what the City is built on. While the panic about technology stocks grabs headlines, it’s largely irrelevant for the wider UK financial services industry.

UK asset managers may manage £13 trillion, but much of this is for overseas clients, and 70% of the shares they manage are listed on foreign exchanges.

Many of London’s real areas of leadership are completely detached from the LSE. For example, the city has a 46% share of the over the counter interest rate derivatives market, hosting $2.6 trillion of transactions a day. London is also responsible for $3.8 trillion a day in forex, twice as much as New York.

It’s not hard to see why. London is perfectly positioned for traders who need to straddle Asian and US market hours.

London has the largest speciality insurance market via Lloyd’s of London, while UK-based insurers manage £1.8 trillion in investment.

You can see this trend play out when you conduct technical analysis (AKA the science of overinterpreting trendlines in a way that affirms your priors).

The biggest single jump in UK financial services’ contribution to the economy in recent years came in the early 2000s.

This period, however, was far from a golden age for the FTSE.

The bursting of the dotcom bubble triggered a slump in many UK stocks, while the strength of the pound against the dollar (the product of lower US interest rates) depressed the international profits of UK companies.

But financial services benefitted from investor uncertainty about the new Euro, the absence of the recession that hit the US and EU at the turn of the millennium, and the growing popularity of cheap mortgages and complex derivatives. We won’t dwell on the last point….

In short, the sector’s growth spurt came from certainty, financial innovation and robust economic growth. The sector hasn’t really grown since then, even as the FTSE recovered - possibly because these have been absent…

The commentary about the LSE is primarily driven by the businesses who do need UK equities to do well to get paid. When you see bananas policy ideas (e.g. the removal of tax-free allowances from overseas shares), the proponents tell on themselves.

These banks and asset managers get an outsized share of voice for a few reasons: i) people who complain tend to get noticed more; ii) the LSE is visible, symbolic target of lazy discourse; iii) hardworking derivatives traders probably don’t regularly ring up journalists to tell them how much money they’re making.

But ultimately, investors and pension funds should be investing in the highest-returning assets, wherever they are. The industry doesn’t have a moral responsibility to prop up UK assets no one else wants.

In fact, the government’s own analysis of its reforms designed to mobilise pension funds finds that greater UK equity allocation would drive no increase in returns, while more exposure to unlisted assets would result in only a 2% uplift over 30 years (with a considerable degree of uncertainty).

“Selling the crown”

If we put the grumbling UK asset managers to one side, I can’t help but wonder if the discussion about the LSE is really just a proxy for something else. A sense that either big companies aren’t staying British or being built in the UK at all.

I have little sympathy for the former concern, but some for the latter.

I can definitely see the argument for not wanting Chinese buyers to control our steel or nuclear industry. These are strategically important assets, and it’s faintly mad that UK governments allowed this to happen in haste, only to repent in leisure.

But when it comes to technology - I’m less sold on the notion that acquisitions from companies based in friendly nations is a bad thing, especially if they maintain a meaningful presence here.

Often the argument seems to boil down to, well, vibes. Companies should have a Team GB stamp on the side and that’s just a good thing. Government should intervene to prevent the stamp from being replaced.

We see this play out most prominently in debates around the sale of ‘UK success stories’, like Arm or DeepMind.

But this argument is usually grounded heavily in hindsight bias.

When Google bought DeepMind in 2014 for £400 million, there was no real controversy around the deal. While DeepMind’s early application of deep reinforcement learning to Atari games was known among AI researchers, the company was largely unknown to the wider world. And it didn’t want to be.

Coverage of the acquisition emphasised the company’s obscurity. For example, TechCrunch noted that:

“But the company may also have to clarify what exactly DeepMind’s AI tech does. The company’s site currently just has a landing page, with a relatively vague description that says DeepMind is “a cutting edge artificial intelligence company” to build general-purpose learning algorithms for simulations, e-commerce, and games.”

There was no clamour to keep DeepMind British until it became highly successful years later.

Before being acquired by Google, the company hadn’t raised a penny from UK investors. This wasn’t because UK investors looked at the company and passed - it’s because DeepMind didn’t pitch them.

Early profiles of the company noted that co-founder and CEO Demis Hassabis specifically decided he wanted to raise money from Peter Thiel. He probably felt that the co-founder of PayPal and Palantir could contribute more than a London-based generalist SaaS VC. How dare he.

And had the government intervened to force DeepMind to ‘stay British’, would the company have turned out for the better? Would it have been able to pursue the same ambitious long-term scientific research? Would it have been able to largely avoid commercial pressures? Demis’s talent notwithstanding, almost certainly not.

Despite having Google stuck next to its name, its team is based in London and makes a vital contribution to the local ecosystem. It supports academic research at UK universities, and members of its team are angel investors or have founded their own businesses. This is exactly what we should want!

Returning to our earlier Arm example - the company was undoubtedly a UK success story. But Softbank bought Arm at a 43% premium on the share price - if UK investors thought it had more potential, they didn’t show it.

When Softbank came to list the company on Nasdaq, no one thought the company merited its now $150 billion market cap. In fact, the consensus opinion among the commentators ahead of the 2023 IPO was that Softbank’s target of $50 billion was too high.

Softbank also helped ARM invest in R&D and significantly grow its headcount, while the company diversified beyond mobile, leading its revenue to increase.

As with DeepMind, the M&A opponents blithely assume the business would have achieved the same success or value independently.

It’s also completely legitimate to ask questions about the company’s current valuation, which is likely being inflated by a couple of points.

Firstly, by investors seeing Arm as a way of getting cheaper AI exposure than NVIDIA offers (even though historically little of the company’s business has been focused on AI…). Secondly, by only listing 10% of its stake in the company, Softbank has introduced a degree of artificial scarcity. This is likely why Arm’s soaring valuation hasn’t been reflected in Softbank’s shares, which currently trade at a 60% discount to the net value of the company’s assets.

With Arm remaining a global company and headquartered in Cambridge, the UK has captured plenty of the benefit. The LSE lost a dollop of listing fees, while UK investors missed out on Arm’s gains. But … that’s their fault! Softbank was prescient and saw more potential than they did, and was rewarded for it. That’s how markets are supposed to work.

The next trillion dollar company?

So if the common examples of British buy-outs don’t tell us anything, should we be concerned about the lack of big, new UK companies?

Probably.

It’s not politically fashionable to say this, but small businesses are not the backbone of the economy. They are more likely to employ outdated management practices, are less competitive, spend less on R&D, and pay less in tax.

A company like Ozempic-maker Novo Nordisk, makes a significant contribution to the Danish economy, because it has been able to achieve scale and win internationally. As a result, they paid $2.3 billion in tax last year and were the primary drivers’ of Denmark’s 2% GDP growth. Their contribution unlocked record spending on defence and infrastructure. Booming exports have led Denmark’s central bank to keep interest rates low, making mortgages more affordable.

It’s an extreme example - but it’s hard to argue that several smaller success pharma companies would have had the same positive effects.

A UK ecosystem full of start-ups that struggle to scale beyond Series A isn’t something to be celebrated.

But I think UK enthusiasts can go too far the other way.

Back in May, then Chancellor Jeremy Hunt described his “yardstick of success” as the emergence of a $1 trillion UK company.

I criticised the framing at the time, and various VCs on X told me to “read the room” and stop hating ambition - but I think the framing’s America-brained and actively unhelpful.

A UK company worth £1 trillion would be worth about a third of the UK’s nominal GDP. Even if such a company were possible, it would pose huge systemic risk. It would hold our entire political system hostage. There are already concerns about Novo Nordisk posing a “Nokia risk” to Denmark - a reference to how the collapse of Nokia’s mobile business tanked the Finnish economy.

The US is a significantly larger economy, with a greater population, more top-tier academic institutions, and deeper markets. I’m not saying the US is necessarily proportionally better on these metrics (although it probably is on a few of them) - but there’s obviously a significant advantage in mass that we can’t replicate.

It is much more reasonable to ask why the UK doesn’t have Novo Nordisk-style companies, with a value in the £250-400 billion range.

I’ll probably return to this in more depth in future, but a few reasons stand out.

Firstly, it’s just harder. I’ll spare you vibes-based cultural anthropology in this post, but if you are building in a smaller economy, your ceiling is lower. By contrast, it is possible to build a successful US-based business that doesn’t export meaningfully.

For example, a company like Robinhood, which has essentially only ever operated in the US up until now, has been able to find close to 25 million customers, and has a market cap of over $30 billion at the time of writing. This isn’t to say Robinhood had it easy or didn’t work hard, but a UK-focused brokerage would have to sign up close to two-thirds of the working-age population if it wanted 25 million customers.

To get to the tens of, let alone the hundreds of, billions - requires the ambition and ability to not only win at home, but to beat homegrown competitors in overseas markets.

If we look at international UK success stories, certain trends emerge.

Revolut, for example, was a beneficiary of the UK’s light-touch regulatory framework for fintech (e.g. around financial data access). The e-money licensing framework enabled fast market entry. If the company had attempted to launch in the US, it would’ve been forced into an arrangement with an established player with a banking license. This may well have involved some form of revenue share agreement.

Other successes, like Arm, Darktrace, or DeepMind reflect the combination of long-standing areas of academic strength in the UK with good entrepreneurs and vast sums of money.

Is there anything we can do to ensure this becomes more common?

Some of it, as investor Ian Hogarth argued in a recent FT essay, does require a better, more ambitious class of investor. The UK government seemingly endlessly subsidises generalist SaaS VC firms irrespective of their performance, but if you’re a solo entrepreneur looking to start a fund? Forget it. None of these schemes are open to you.

In other cases, it’s a case of the government getting out of the way. This is part of the Revolut lesson.

It’s common for champions of Europe to argue that US taxes or regulations are as onerous or worse.

And it’s true. The US enforces antitrust aggressively, and capital gains tax in California is higher than it is in the UK.

But remember the scale point. The US can get away with being bad on these questions, because it’s the US. When your market is smaller, the bar is higher. You should assume talent will want to move to the bigger market, because of its existing dominance, networks, and capital. That means we have to fight hard to keep it here! Entrepreneurs need to think “I’m building X, my ambition will be to scale this globally, but it’s easier to start it in the UK”. But there’s no point getting upset with the LSE not competing with Nasdaq - of course it doesn’t!

The key for this will lie in structural or regulatory reform. Not in more subsidy schemes. It’s also likely that many of these reforms would be politically difficult to swallow.

Potential measures like deregulation, lower taxes, and reduced employment protections are all highly contentious and come with trade-offs. There’s also no guarantee that they would work. But I’m sceptical that there’s a magic, pain-free policy lever that we can pull that simultaneously preserves the UK’s current economic model while creating a new generation of mega-companies.

Unfortunately, successive UK governments have instead presided over a stagnant economy and political chaos, while scratching their heads about people making rational business decisions.

So, should we save the LSE?

We started the piece with the LSE, before proceeding to declare it irrelevant. Does the UK’s declining stock exchange have a role to play in all this?

The Tony Blair Institute and Onward recently published a report calling for UK capital markets reform, which recommends closing down the AIM, reducing the burdens on small fund managers, and making listing cheaper and quicker for companies.

These measures are likely all sensible and I see limited downside in pursuing them. At the same time, I suspect they’re only going to make a marginal difference to the LSE. In our new world where companies start in Britain and expand globally, I suspect the majority would still rather list on the NYSE or the Nasdaq, and that’s fine.

But, if the LSE doesn’t play host to the next £250 billion company, the kind of reforms the TBI outlined might help it play a different role. We can probably learn something from Sweden’s example. Swedish global success stories like Spotify are listed in the US, but their domestic exchange acts as a source of capital for smaller businesses. In Sweden, it’s cheap and easy to list a company, you have a willing and motivated retail investor community, and competent asset managers who aren’t snooty about luring small to mid-sized companies from abroad.

As I said in my defence piece a couple of weeks ago, officially accepting the UK isn’t always going to be ‘world-leading’ at everything isn’t good for pride - but it would make designing good policy significantly easier.

Disclaimer: These are my views and my views alone. They aren’t those of my employer (Air Street Capital), whoever manages my pension, or anyone happens to live in London. This is not financial advice. Don’t arrange your finances based on strangers’ Substacks. Somewhat damningly, I hold no shares in individual LSE-listed companies, but am invested in a UK index fund, while Lindsell Train used to charge me high fees to deliver multiple quarters of benchmark underperformance. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Sir Humphrey : Bernard, what is the purpose of our defence policy?

Bernard Woolley : To defend Britain.

Sir Humphrey : No, Bernard. It is to make people believe Britain is defended.

Bernard Woolley : The Russians?

Sir Humphrey : Not the Russians, the British! The Russians know it's not.

Yes Prime Minister (1988)

Introduction

As part of my day job, I’ve written and co-written pretty extensively about UK defence. I’ve taken a particular interest in the broken world of Ministry of Defence (MOD) procurement. 

For decades, the UK has overpaid defence primes for underpowered, bespoke systems that arrive years late and significantly over-budget. This includes £1.5 billion on drones that couldn’t fly in bad weather, £5.5 billion on armoured vehicles that injured their operators, and £3.2 billion on a battlefield communications system running half a decade behind schedule.

In my interactions with ‘stakeholders’ - whether they’re civil servants, military, or executives at defence companies - I routinely hear some version of “we’ll sort it out when there’s a war on”. The idea being that the system just needs a short, sharp shock to get its act together.

This dangerous assumption is implicit in a couple of things. 

Firstly, the government’s repeated insistence that it will increase defence spending to 2.5% of GDP “as soon as resources allow”. This assumes that the outside world will arrange itself to the Treasury’s timetable. 

Secondly, last week, the UK government announced plans to cut five warships, a fleet of drones, and to retire dozens of its oldest Chinook helicopters early. This will save approximately £500 million over the next five years.

Some of these cuts are reasonable, including Watchkeeper (the drone programme referenced above), but others are more concerning. For example, the Chinook replacements won’t start arriving until 2027. Nor will the Navy’s new class of frigate. The decision to retire amphibious assault ships HMS Albion and HMS Bulwark had been considered by the previous government. Then Shadow, now real Defence Secretary John Healey said earlier this year that the proposals he is now enacting “would further hollow out our forces and raise serious concern over future operations for the Royal Marines”.

These decisions have been taken in the hope that we can muddle our way through until new capabilities arrive and in the event of a crisis, something will turn up.

But I believe that the idea that UK defence can ramp up quickly in an emergency is profoundly short-sighted. It involves turning a blind eye to the depleted state of our armed forces, overestimates how easy a sudden ramp-up would be, and relies on a selective reading of history.

The cupboard is bare

In September of this year, the Lords International Relations and Defence Committee published a devastating report on the state of UK defence readiness. The committee found that “the UK’s Armed Forces lack the mass, resilience, and internal coherence necessary to maintain a deterrent effect and respond effectively to prolonged and high-intensity warfare”. 

They point to the UK’s inability to simultaneously deploy fighting forces, while guarding long lines of communication and protecting critical national infrastructure. This stems from our forces’ size and structure being “predicated on the belief that conflicts would be resolved within weeks, rather than years”.

The report concludes that “governments, have so far, lacked an honest narrative about Defence’s ambitions, resources, threats and risk”.

You see this everywhere.

The armed forces have shrunk by nearly 30% since 2000.

Even recent shows of strength, such as our strikes on Houthi targets in Yemen earlier this year, highlight underlying weaknesses.

The UK’s contribution was largely symbolic, four Eurofighter Typhoons striking two targets. Delivery of the second and third squadron of F-35B Joint Strike fighters has been delayed. HMS Diamond, the Type 45 destroyer currently protecting shipping in the Red Sea currently can’t attack land targets. No submarines were available, due to the delays in operationalising the new Astute-class submarines. In fact, Britain’s current submarine fleet is barely large enough to protect the nuclear deterrent.

Despite drone warfare gaining in salience, the UK is a laggard. The Watchkeeper programme was a shambles, while the ‘Defence Drone Strategy’ published earlier this year was primarily a set of pictures, case studies, and aspirations to partner with industry. It was revealed earlier this year that the UK’s drone trial squadron had not conducted any trials since being formed in 2020 due to a lack of drones.

Meanwhile, the Defence Artificial Intelligence Strategy seemingly envisages the UK prioritising the use of AI for career management over enabling swarms of drones.

At the end of last year, the National Audit Office concluded the MOD’s Equipment Plan for the next ten years has a £16.9 billion black hole centre, while the MOD can’t keep track of the equipment it does have.

The rot runs deep

It would be easy to attribute the poor state of critical institutions to an extended period of austerity following the 2008 global financial crisis. There’s definitely some truth in this - but it’s incomplete as an explanation. 

We’re witnessing the consequences of a 30-year trend of borrowing from defence to pay for the rest of the public realm.

This trend started in the 1990s after the Cold War, but the UK didn’t update its approach as the War on Terror broke out.

You’d be forgiven for assuming that the era of deep UK-US cooperation with simultaneous major deployments would be a golden age for UK defence spending. 

But the mid-2000s were counterintuitively an era of retrenchment for UK defence. 

Much of the cost of the Iraq and Afghanistan deployments was met via a special Treasury contingency rather than through the core MOD budget, which was squeezed. In this era, ministers could confidently state that defence receiving real-terms spending increases. But these were often a sliver above low inflation rates and did little to reduce pressure on resources. Even as British forces were deployed overseas, they were the target of penny-pinching.

In December 2003, the government published a defence white paper. It proposed a move away from heavy-weight land forces, and outlined reductions in equipment and manpower. This was followed by a sequel in July 2004 that spelt out the specifics, including a 10,000 fewer personnel across the three services, reductions in numbers of helicopters, missile launchers, nuclear attack submarines, patrol vessels, and frigates. 

These reductions would be offset by the embrace of “network-enabled capability” where, thanks to technology, we would be “able to exploit effects-based planning operations, using forces which are truly adaptable, capable of even greater levels of precision and rapidly deployable”. In essence, shortening time for decision makers and between shooters and sensors.

This would in turn support “effects-based operations” - military activities designed to create specific psychological, physical, or behavioural outcomes, rather than simply destroying targets. The MOD stopped short of defining what this meant specifically in Iraq or Afghanistan. 

Hindsight may be 20:20, but there were plenty of voices expressing alarm about this approach at the time. 

In June 2004, the House of Commons Defence Committee published a highly critical report on the 2003 white paper. 

It warned that the MOD was abusing the language of doctrine, stating that: “The suspicion has grown that the focus on agility, effect without mass and the move away from a platform focus has less to do with an intellectually coherent strategy of effects-based warfare than with the need to ‘cut our cloth’ as best we can”. Elsewhere the report called on the MOD to “urgently” tackle ‘gapping’ - the practice of leaving non-essential posts vacant to tackle manpower shortages.

As with the current government, capabilities were scrapped, new ones were promised, but without budgets or timelines for procurement.

At the time, Michael Codner, a senior defence analyst at RUSI noted that the plan was “published at a time when there was widespread awareness that there was not enough money in the Defence Budget to fund defence activities and the equipment plan and the imbalance was too large to be addressed by modest increases in defence spending and greater efficiencies”.

There was good reason to be concerned. 

Since 1998, the UK had planned on the basis that while its armed forces could be engaged in two substantial deployments simultaneously, these could not run concurrently for more than six months. They also did not expect both deployments to “involve warfighting”. 

By 2004, the UK had comprehensively blown through these assumptions.

While the US and UK governments were publicly maintaining the pretence that Iraqi self-governance was just around the corner, this was beginning to collide with reality. By the time of the 2004 capabilities paper, attacks by insurgents were already gaining in sophistication, Muqtada al-Sadr's Mahdi Army uprising was under way, and we’d seen the First Battle of Fallujah. Drawdowns weren’t imminent.

Before we’d exceeded these assumptions, the UK’s armed forces were already stretched. The official Iraq Inquiry spells out the state of pre-invasion preparation in black and white.

The report concludes that “after the invasion began, it became clear that some personnel had not been equipped with desert clothing and body armour … and there were shortages of ammunition”. As well as actual shortages - poor asset tracking made it hard to get equipment to the right people. This deficiency had been identified during the 1991 Gulf War, but the MOD had not remedied it 12 years later. 

Sir John Reith, former Deputy Supreme Allied Commander Europe within NATO described the equipment policy as “just enough just in time” - and said that it didn’t work.

During the long counter-insurgency phase of the operation, one vehicle came to symbolise the deficiencies of the UK’s approach: the Snatch Land Rover. 

These lightly armoured patrol vehicles had been designed for use in Northern Ireland to protect against small arms fire and basic explosive devices during the Troubles. They were inappropriate against larger IEDs and rocket propelled grenades, but as General Sir Mike Jackson, Chief of the Defence Staff from 2003 to 2006 told the inquiry: 

“Snatch Land Rovers were deployed to Iraq because they were available or could be made available as we drew down from Northern Ireland, and without them, it would have been completely soft-skinned Land Rovers. That’s where the state of the equipment inventory was at that point”. 

37 British personnel died in these vehicles in Afghanistan and Iraq, leading them to be dubbed ‘mobile coffins’.

Anyone who followed the news regularly in the UK from about 2006 and 2010 will remember the widely criticised shortage of Chinook helicopters in Afghanistan. 

2006 marked the deployment of British forces to Helmand Province to shore up the Afghan authorities, who were struggling against a resurgent Taliban. 

While the military’s senior leadership assured the government that a 3,000-strong British task force could get the job done, UK forces were undermanned and underequipped. Theo Farrell’s scathing account of the war, Unwinnable quotes the reactions of members of the 16 Air Assault Brigade, after they were informed that the taskforce could not exceed 3,150. Experienced officers warned the mission would be a disaster, and identified “significant shortfalls in core equipment areas such as helicopters” . But in response, “the consistent line was that we had to make do”, as multiple requests for additional equipment were denied.

The deteriorating relationship between the Chancellor and the Prime Minister meant the Treasury refused to make more resources available, and there was little pushback at the most senior military level.

Both the military and political factions appeared unbothered by the several years-long overstretch of their own planning assumptions. It’s not for nothing that Farell concludes that: “‘Ultimately, the British campaign in Helmand was characterised by political absenteeism and military hubris’.

The Iraq Inquiry also flagged up a similar failing among military leaders, finding that a “can do” attitude in the armed forces led to “over-optimistic assessments” at the expense of “accurate and frank reporting”. 

To summarise, even when UK defence spending was higher, the UK was unable to adequately resource military activity that it had planned in advance. The notion, with the armoury empty, that it would be able to resource unplanned activity seems Pollyannaish in the extreme. 

Things can only get worse

When the 2010 Conservative-Lib Dem Coalition came to power and embarked on a course of austerity, Sir Rodric Braithwaite, former chairman of the Joint Intelligence Committee was scathing about their inheritance, saying:

“The new Coalition government came to power to find that their predecessors had bequeathed them a national defence strategy that was intellectually void; a military procurement policy paid for on the Micawber principle - that something would surely turn up, but measured in billions rather than sixpences; a bunch of generals on the verge of revolt; a vision of Britain’s place and influence in the world based largely on wishful thinking; and a horrendous financial crisis.”

The 2010 Strategic Defence Review introduced significant cuts in personnel across the services, a 40% reduction in Britain’s fleet of Challenger 2 tanks, the culling of the Navy’s fleet of Harrier jump jets, the decommissioning of 9 frigates and destroyers, the mothballing of aircraft carrier HMS Ark Royal, and the scraping of the Nimrod maritime patrol aircraft shortly before its completion.

This stripping of the armoury created operational problems in the coming years. 

For example, the Harrier was the UK’s only fixed-wing aircraft capable of operating from a carrier. Its early retirement left the Royal Navy without this capability for over a decade until the F-35B Lighting II became operational. This meant that during the UK’s operations in Libya, the UK had to pick the more expensive and less flexible option of operating from Italy, rather than from a nearby carrier.

Scrapping the Nimrod left the UK without a maritime patrol aircraft - a useful capability for a geopolitically significant island nation. The Nimrod replacement, the P-8 Poseidon, wasn’t delivered until 2019. In the meantime, UK personnel had to be loaned to allied nations to maintain their skills. This left the UK reliant on US aircraft and a diminished frigate fleet to spot and counter Russian submarine incursions - which analysts warned about at the time. Over the course of the decade, Russian incursions to map out undersea cables increased.

Plucky British improvisation?

Part of our national mythos is this idea that Britain can always pull it together once its back is against the wall. Ingenuity, improvisation, and spirit will lead us to triumph against the odds.

Let’s break the first rule of arguing on the Internet and look at World War II.

The opening phases of the war, with defeat in Norway through to the fall of France in June 1940 were not Britain’s finest hour. Interwar intellectual stagnation in the military and an army constructed on the basis of what politicians believed the country would tolerate paying for was disastrous. There was a shortage of tanks, defence planners seemed uninterested in them, and little consideration had been given to the style of land-air operations that the Germans would use to devastating effect.1

But despite this early military incompetence - the UK industrial-base remained resilient. So much so, that Churchill later admitted to being overly pessimistic about the state of the country in his memoirs. By the outbreak of war, Britain was both self-sufficient in energy and the world’s biggest exporter of coal. It was also the world’s biggest exporter of arms. Between 1935 and 1939, arms expenditure increased at a faster rate than it did during the pre-World War I arms race. During the interwar years, the Royal Navy out-built every other navy in the world in almost every class of warship. 2

In short, it’s easier to compensate for doctrinal screw-up when the arsenal isn’t empty and you’re capable of replenishing at least some of it.

If we want to continue on the historic detour, Britain in part prevailed in the Napoleonic Wars of the nineteenth century thanks to the significant investment in its naval capacity in the 1780s and 90s. Instead of cutting spending following the American Revolutionary War of 1783, as was normal in the eighteenth century, William Pitt’s government increased it. The number of men at dockyards was maintained at wartime levels. Between 1783 and 1793 alone, 138 ships of the line were built or given major repairs. This was done while paying down the national debt.3

When you’re working a smaller defence-industrial base, as the UK is now, that’s been starved of orders for a long time, you can’t just flick a switch and fire up production. 

Producing weapons or ammunition requires staff, materials, equipment, and infrastructure. As European governments ramped up orders following the Ukraine invasion, they found 10-20 month waits for unguided shells, rising to 36 for guided ones. With concerns that the UK would exhaust its stocks of ammunition in 8-10 days in the event of intense warfighting, this isn’t ideal. 

Rheinmetall, the German arms giant, is on a buying spree to help it meet orders. In August alone, it acquired a South African engineering firm and an American vehicle manufacturer. One month earlier, a new factory in Hungary was completed. In February, it acquired a majority stake in a Romanian vehicle manufacturer. 

Parallel efforts are not happening on a similar scale in the UK. This isn’t because Rheinmetall is particularly smart or patriotic, while their British competitors are stupid. It’s because UK-based manufacturers do not believe that they will receive the orders necessary to justify this kind of outlay. 

When disaster strikes

It’s of course impossible to know exactly how British defence would respond in the event of an emergency, But there have been recent incidents where crumbling arms of the state have been tested. And they performed dismally. 

The recent Coronavirus pandemic was handled poorly in many countries, but the UK experienced one of the worst death tolls of large European economies. Save for the intervention of a few talented outsiders, it could have been worse.

While the official inquiry into the pandemic is still going on, Module One, focused on pandemic preparedness, has been published. They found that despite “reams of documentation” and a “labyrinthine” institutional setup, the UK’s pandemic preparedness was damaged by bureaucracy, groupthink, outdated assumptions, and a lack of interest or challenge from ministers. Sound familiar?

The authors had “no hesitation in concluding that the processes, planning and policy of the civil contingency structures within the UK government and devolved administrations and civil services failed their citizens”.

The UK’s Vaccine Taskforce (VTF) was one of the few bright spots of an otherwise shambolic pandemic response. 

Chaired by venture capitalist Kate Bingham, the VTF was tasked with rapidly sourcing and deploying a Covid vaccine. It operated with a high degree of autonomy and made quick decisions on procurement and investment, backing multiple different vaccine candidates before it knew which would succeed. This meant that the UK became one of the first countries to begin mass vaccination and achieved high vaccination rates relatively quickly.

But the bureaucracy was seemingly determined to kill off this agile, risk-tolerant body from day one.

The frustration is palpable in the pages of The Long Shot, Kate Bingham’s book about the experience. She describes her team having to waste time filling in hundreds of pages of Whitehall business case documentation:

The standard Whitehall case required us to quantify the benefits to the UK economy based on the extent to which vaccines put a stop to the pandemic - also impossible to determine in June 2020. [...] We needed to produce estimates of where the economy would be without a vaccine and how much larger it might be following vaccination, but without yet knowing how effective the vaccine was at saving the lives of different age groups, or who would be vaccinated and when. A graph was inserted for the sake of form.

Within two months of the taskforce being launched, the National Audit Office began investigating the VTF to ensure it used “their resources efficiently, effectively and with economy”, jamming up its work with requests for information and meetings. Meanwhile, Kate Bingham herself was the target of vicious partisan attacks in the media.

In short, the UK’s bureaucratic and political class were keen to kill off the most successful part of our response. We’re very lucky they didn’t succeed. 

But Ukraine?

As Russian tanks rolled into Ukraine in February 2022, most observers expected that Putin would quickly achieve his war aims. Famously, the US offered to airlift President Zelensky to safety. The Ukrainians’ early and ongoing success in holding out against a numerically superior military has rightly been attributed to the ingenious use of asymmetric tactics, supported by the total mobilisation of society in support of the war effort. 

One of the most visible manifestations was the explosion of teams working on drones. These efforts have now evolved into a highly capable industry that’s respected internationally. Can we take hope from this?

Yes and no.

While there is much we can learn from the admirable speed and efficiency with which the Ukrainians now procure technology, societal mobilisation carries fewer transferable lessons.

I don’t want to stare into the crystal ball too much, but I think the chances of Russian tanks rolling past Watford Gap service station are slim. 

In the event of an incursion into Finland, an attack on one of the Baltic States, or an escalation in Moldova - the average British citizen will not be fighting for their survival. I doubt there will be the political will to mobilise all of society in response. In reality, the UK is on course to be a junior partner in a multinational coalition. 

I sometimes hear that the UK’s strength is its technology or special forces. These are more valuable than mass. This may well be true, even if there’s scant evidence of the former in defence. But when a country like Estonia, Poland, or Finland is putting a meaningful proportion of its adult population on the line, while the UK can’t deploy equipment or people at scale, can we reasonably expect to have an equal seat at the table? 

In the worst case scenario, a future conflict with Russia could overlap with a confrontation in the South China Sea. A recent RUSI report noted that in the event of a crisis in the Taiwan Strait, European nations risked facing critical gaps in missile defence and anti-submarine warfare. It’ll be all hands to the pump.

The isolationist will say that this doesn’t matter. What happens in Chișinău doesn’t matter in Crewe. But this is where sangfroid can tip over into being sans brain.

Even if we put morality aside, a Russian-dominated or conflict-ridden Central and Eastern Europe would be dire for UK interests. Bilateral trade with Poland alone is worth £30 billion a year. Major UK companies have significant market presences in these countries, while Eastern Europe is deeply integrated into UK manufacturing supply chains for automotive parts, electrical components, and agriculture products.

Forfeiting our seat at the table leaves the UK dependent on others for the defence of critical national interests.

Whither Britannia? 

Preparing for war in the same way we prepared for the pandemic isn’t an option.

So where does this leave us? 

For a start, urgent increases in defence spending are needed. We should be moving towards 3% of GDP urgently at a minimum.

As we move to restock the armoury, important questions need to be asked about the exact role the government sees the UK playing in any future conflict. Historically, the MOD has ducked these questions, leaving us with ‘a shopfront military’, in the words of one former senior official. It means we’ve spent close to £7 billion building two aircraft carriers to project force, while draining the resources needed for the accompanying planes and escorts.

Do we envisage a world in which the UK plays a meaningful role in the defence of our European allies, while we send a carrier group to the South China Sea as part of an ‘Indo-Pacific tilt’? Or should we scale back our focus to Europe and the Mediterranean?

There are some things that we definitely need to do. For a start, greater mass is essential. The vision of the UK as some kind of elite defence R&D island is appealing, but fanciful. 

Whatever we choose to do, our armed forces will need to be larger and we will have to pay for them. 

But there are real questions and trade-offs about the extent to which we balance older-school mass with new technology. While the drone has become the defining image of the war in Ukraine, much of the conflict still resembles an early twentieth century clash of mass armies. 

We also need to have an honest conversation with both ourselves and our US and European allies about the role we will and won’t play in any future confrontations with our adversaries. The official acceptance that we are a second-tier military power will be bad for our image, but significantly better for our ability to plan. 

But wherever we land on these questions, business as usual with more money is not going to be the answer. The MOD’s current procurement system, for example, is beyond repair. While I’m usually sceptical of new units or agents, there’s a strong case for moving defence acquisition out of a captured MOD bureaucracy. 

We could bring outsiders in to operate it, exempted from civil service payscales and run competitive trials for solutions as opposed to a paper requirements process. This would also free the process from the MOD’s fixation on bespoke solutions. Rather than overpaying contractors to format a special GB drone, we might just buy the one that works quickly and in bulk.

By commissioning a Strategic Defence Review led by credible, independent outsiders, the new government has given itself the space to begin having these difficult discussions. In October, the government created a new National Armaments Director to drive procurement reform. An accountable individual is welcome, but it remains to be seen if they’ll have the powers they need to bring about change.

We can only hope so. 

More MOD ‘strategies’ that make vague allusions to Britain’s place in a changing world won’t cut it.

Subscribe now

Disclaimer: These are my views and my views alone. They aren’t those of my employer, any current or former UK Government officials, or anyone I’ve spoken to who’s served in the armed forces. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Photo credits:

Canal Hotel Bombing: MSGT JAMES M. BOWMAN, USAF, Public domain, via Wikimedia Commons.

HMS Queen Elizabeth and HMS Prince of Wales meet at sea for the first time. Petty Officer Photographer Jay Allen, OGL 3, via Wikimedia Commons.

Introduction

The UK is proud of its tech ecosystem. You’ll often hear its champions boast about the quality of our universities and R&D base, or the presence of tech companies like Google DeepMind in King’s Cross. But if you spend enough time talking to civil servants and venture capitalists, you’ll also hear two acronyms: EIS and VCT.

EIS (the Enterprise Investment Scheme) and VCT (Venture Capital Trusts) were created by the UK government in 1994 and 1995 respectively and have been renewed ever since. These schemes were designed to incentivise wealthy individuals to invest in start-ups at a time when early-stage capital was scarce in the UK. By giving investors generous tax relief and downside protection, this was meant to de-risk investment in otherwise high-risk enterprises. 

Subscribe now

While the schemes have chopped and changed, the basic idea has stayed the same.

For EIS, investors can claim up to 30% of their investment (up to the value of £1M, or £2M if at least £1M is invested in “knowledge-intensive companies”) in qualifying businesses as income tax relief either in the current or previous tax year. They also receive a capital gains exemption when they sell their investment provided they’ve held it for at least three years, and loss relief against income tax. Shares bought via EIS are also inheritance tax exempt.1

VCT investors buy into the public offering of a listed company that holds shares in unlisted companies or AIM stocks. They offer a similar deal on income tax to EIS (albeit with a £200k ceiling) and with a five year year holding period. No capital gains tax is levelled on dividends.

One of the new government’s first acts in office was to extend these schemes to 2035. In a given year, around £2B is raised via EIS and around £1B through VCT.

So… great news for founders?

Unfortunately, EIS and VCT are a case study in the danger of tax breaks and subsidies. Tax relief that was once meant to act as a derisking sweetener has become the primary purpose of these schemes, with significant distorting effects. They also show us the peril of governments outsourcing the monitoring of policy to ‘stakeholders’ and how, if something is dressed up as ‘supporting business’, it’s possible to get away with murder.

High fees, poor performance

I’m going to start with one caveat. While some individuals choose to make EIS investments, many choose to invest in EIS funds. Essentially, you put your money in and over the course of (usually) a tax year, it’s invested in a portfolio of EIS-qualifying companies. 

This works well for many EIS investors, who don’t want source deals or manage investments in individual companies. These funds, rather than individual investments, are going to be my focus here, although I will also briefly touch on angel syndicates - where groups of EIS-eligible investors come together to pool their resources.

If we want to assess how well a scheme is working, the performance of the biggest investors in EIS-qualifying companies is a good starting point. EIS is a notoriously untransparent industry, but the Tax Efficient Review (says it all) produces performance breakdowns, which some funds use in their marketing. I’ve lifted this table from Mercia’s 2023 materials.

EIS funds tend to launch a new fund every year, so these stats aren’t year-on-year rises and falls. Each number reflects a different fund vintage and how the investments made in that specific tax year have fared.

These figures are … underwhelming. Risky, illiquid investments are meant to be rewarded with outsized returns. But more importantly, these numbers don’t account for fees, which are meaningfully higher than in standard VC. 

In addition to the typical 2% management fee and 20% carried interest, EIS layers in a world of other charges. These include an initial charge for ‘setting up the portfolio’ (up to 5-6%) and an annual administration fee (0.2-0.5%). Most funds have a hurdle rate (a minimum level of performance above 1x before the performance fee is deducted), but there are several exceptions. Mercia, for example, doesn't have one.

There are also the fees on investee companies - which we’ll get to in a little bit…

These fees begin to stack up. The TER folk have simulated fees on a portfolio of £100k invested for five years, assuming 20% annual growth on all investee companies, no write-offs, and with all companies sold together after five years. This, of course, is a highly unrealistic scenario, but it reveals how quickly the fees stack up. 

Considering the performance of many of these funds, it’s possible that the performance fee could go down, but the cost per pound of profit increases, due to the application of fixed annual fees and VAT.

This means you need to adjust all of the numbers in the nice colourful table down significantly. 

To be as generous as possible, we’ll pick the best-performing recent year in the table - 2019/20. Assuming standard EIS fees but without applying the income tax relief - for investors at the time this table was compiled in 2023 - a fund needed to make 1.2x to break even. For the the annual return to cross 5%, it needed to clear 1.3x

If we look at a benchmark for non-US PE and VC firms from 2023, convert the EIS funds’ gross multiples to annualised returns net of fees, 4 of the 18 funds clear the bar, while the rest underperform, significantly for the most part. (I accept that trying to convert EIS returns like this is pretty crude, due to differences in reporting, but it’s the best I can do with the available data.)

Are the poor returns just a product of EIS? 

VCT fees tend to come in a percentage point or two lower, but so do the returns. (Ignore the highlights, this table is from Mobeus’ marketing materials.)

Tax relief isn’t derisking these investments, it’s keeping them alive.

Bad incentives = bad outcomes 

If you accept that the purpose of EIS is really to act as tax relief and you ignore most fund managers’ rhetoric around innovation, the lacklustre performance makes a lot more sense. And to be fair to managers, it’s pretty transparent.

You see it in the advertising:

Or in the other services they provide:

If you’re an investment manager in the tax efficiency game, you have two primary incentives: (i) to get as much investors’ money out of the door as possible during a tax year, (ii) to protect your downside. 

In short, it’s a perfect mechanism for allocating capital inefficiently. 

For most EIS/VCT funds, this manifests in poor company selection or a bias towards doubling down on their existing portfolio, regardless of growth prospects. In fact, one founder I spoke to told me that their rejection from an EIS fund explicitly said that the technology they were building was “high risk, high reward”, so it was incompatible with the fund’s thesis.

This rejection makes no sense in venture. If what you’re selling is 30% income tax relief, of course you’d rather choose companies that reliably provide a 1-1.5x return, so you can just give people their money back later.

You can see this when you go into the nitty-gritty of their year-on-year investments. Unfortunately, I don’t have an army of research assistants, so I’m going off the limited number of freely available TER fund breakdowns I could access.

The first product I looked at was Mercia EIS Fund and Knowledge-Intensive EIS Fund.

I noticed a few companies that were ‘complete write-offs’ despite the fund having doubled down on them year after year. Sometimes a business seems promising and doesn’t work out. But what I found striking was that even at the time, many of these companies were displaying limited signs of life.

For example, over four vintages they invested in a company called Rinicare, which offered technology for remote patient monitoring. 

Far from being a young, innovative company, Rinicare had existed for 17 years before Mercia made their first investment in 2019. In that time, it had failed to grow or achieve any meaningful commercial traction. It largely appeared to live off government grants. Within a year of investing, the team had shrunk from 12 to 11, and Mercia doubled down again after it declined to 8.2 

The business hired back a couple of people the next year, but it’s hard to look at a team struggling to stay in double digits and significant losses 17-19 years in, and understand what the Mercia team saw in the business. 

Rincare isn’t an isolated example, Manchester Imaging and Pertinax Pharma are two university spinouts Mercia repeatedly invested in before a complete write-down. If you follow the journey through their accounts, you again see more and more money going in, even as growth (by basically any metric) remains elusive.

You don’t just see evidence of zombie-fication in the write-downs. 

If we look at the Guinness EIS Fund,3 you can see plenty of examples of seemingly random companies that have received money repeatedly, despite limited growth in valuation. This might be because the fund is marketed as engaging in “risk mitigation” despite the fact that the government explicitly says EIS is designed to support “high risk” companies. It also explains their focus on leisure, food, and retail businesses. All noble sectors, but worthy beneficiaries of tax breaks for high-risk companies?

It’s an open secret in London that if you’re struggling to raise money from conventional VC funds, you approach EIS funds in February. Their keenness to invest leftover cash means the due diligence will usually be flimsy. Towards the end of the tax year, many will also solicit their existing portfolio, asking if they’d like to take more money (whether they need it or not). A friend of mine spoke to a manager at one of the funds in the returns table in March last year. My friend asked how they were doing and the manager replied that they were super-busy “getting money out the door”.

Free money for founders, what’s not to like? A few things.

Firstly, zombie companies are bad. 

A lot of start-ups fail, and they should be allowed to. That’s how talent and ideas are recycled back into an ecosystem. Politicians repeat the myth that small businesses are the backbone of the economy, but they’re usually less productive and spend less on R&D. Incentivising the creation of new small businesses that don’t scale isn’t good in and of itself.

Secondly, if you’re a founder focused on building a killer business, a partner focused on downside protection doesn’t have the same incentives as you. Their instinct will be to push you to market before you’re ready, steer you away from ambitious R&D, stuff cash into your business and dilute you on their timelines, or to replace you at the first sign of difficulty. 

Downside protection also includes a bunch of other annoying dynamics. 

A founder I spoke to who received their pre-seed funding from an angel syndicate told me that as soon as the three year EIS holding period expired, he was besieged by investors demanding that he open up secondary share sales. Because of the vagaries of their tax management, these investors preferred to cash out quickly, rather than stick around and see the value of their investment increase 8x in an upcoming funding round.

If you look at the Guinness table, you see clear evidence of tranche funding - common in EIS world. Regarded as bad practice in VC much of the time, dicing up money into tranches adds operational complexity, decreases founder leverage, and creates misaligned incentives. Depending on the exact details, it can create a world where founders have to focus on ‘milestone theatre’ - where resources are diverted into short-term ‘wins’ at the expense of the long-term plan.

VCTs come with all of the problems above and more. Alongside the tax breaks, VCT investors are driven by dividends. These dividends have to come from somewhere - either portfolio companies being encouraged to burn through their cash to pay them out or via a sale. Again, it’s not hard to see the mismatch between the fund’s short-term incentives and a founder’s long-term ambitions. 

Money with menaces

My little tour through company accounts didn’t just raise questions about EIS fund due diligence procedures, it flagged up another unedifying element of the industry: the fees. 

EIS fund managers are partly compensated on the basis of fund performance, but the funds often … don’t perform that well. This means the money has to come from somewhere else: the portfolio. Not only do EIS investees have to hand over equity in exchange for investment, they also have to part with cash. 

There are some funds who, to their credit, don’t do this (shout out to Molten, MMC, and Octopus), but some of the terms on offer, elsewhere, seem distinctly founder unfriendly. In some cases, funds split fees evenly between investees and investors. 

For example, Deepbridge advertises that investors don’t pay fees on subscription. Instead, founders pay.

Deepbridge isn’t alone. For example, over at Guinness Ventures, founders get these terms.

The picture isn’t much better in the VCT world. 

If you tried to offer these terms in the Valley, you would be met with a mixture of astonishment, derision, and fury. 

Am I just cherry-picking a few bad examples? A Europe-wide survey of term sheets called out this practice specifically in the context of EIS and VCT funds, precisely because it’s unusual.

And EIS and VCT over-representation in the high fee bracket is stark.

The same founder I spoke to above saw their angel syndicate charge 5% upfront. One member of the syndicate pushed them hard to use a specific lawyer (a friend of theirs) to manage the round, who charged exorbitant fees. When the founder complained about the fees to the syndicate, they were told: “If you think these fees are high, wait until you have to pay private school fees”. The level of empathy any founder wants from their investors.

“We asked some investors if they liked tax breaks and they said…”

Considering billions of pounds have flowed through EIS and VCT since their inception, you’d assume the government would be keen to understand how well they worked. Right?

Wrong.

Despite tens of billions of pounds having been invested through these schemes, the government has been … incurious. The closest thing we have to a recent evaluation came from HMRC in 2022. This ‘analysis’ is a perfect encapsulation of the “We asked people with a vested interest if they like free stuff” archetype of policy work Anastasia and I described a couple of weeks ago.

The evaluation primarily consists of asking people with tax breaks if they like tax breaks (yes they do) and businesses dependent on said investors’ money if they would rather be insolvent (no they wouldn’t). Other questions include HMRC asking fund managers if they believe they perform rigorous due diligence (yes) and investee companies if their work is innovative (also yes). Surprise!

This bracing interrogation did reveal a few interesting tidbits. Namely, that the primary motivation for investors is indeed the tax incentive and that the majority of the recipients admit that their businesses would have struggled to keep the lights on without the support. 

If a business is unable to keep the lights on without money provided as a result of tax incentives (and on potentially disadvantageous terms), this is likely a reflection on company quality. At least in some cases. HMRC does not consider this scenario at all and instead frames these results as an example of policy correcting a ‘market failure’ and asserts (with no evidence) that: “Unregulated free markets will likely under-produce innovation.”

The evaluation also uncritically repeats a series of complaints from founders about VC. These include that i) it’s inaccessible if you’re not based in London, ii) small cheque sizes aren’t available, and iii) all investors want immediate significant dilution. All claims that can be judged with empirical evidence (and are incidentally all untrue), but HMRC decided not to check.

What’s totally missing is any assessment of EIS/VCT recipients’ actual contribution to the UK economy or whether the scheme represents value for money to the taxpayer.

Taking a step back

If the government isn’t going to take the time to analyse this work seriously, has anyone else?

There isn’t a stack of academic evidence and much of it is quite old. One 2009 study went as far as describing these schemes as “largely an act of faith by governments” and cautioned that insufficient work had gone into developing evaluation methodologies. Their warning went unheeded.

While EIS remains hard to assess, there is slightly more academic work around VCTs. This largely underscores the self-defeating elements of the scheme’s design.

For a start, the VCT market is stunted by low levels of liquidity. The rationale for making VCTs public limited companies (rather than real trusts) was to reduce the risk of holding unquoted assets by creating a secondary market with liquidity. 

But as one 2011 study observes, there is no meaningful secondary market for VCT shares. By not extending VCT tax advantages to secondary purchasers, trading volumes remain low and a meaningful bid-ask spread (the difference between the highest price a buyer is willing to pay and the lowest price a seller is willing to accept) opens up during the holding period. These range from 6% up to more than 30%. With most commonly traded shares, you’d expect less than 1%. Considering the small returns many investors make on VCTs, even a 6% spread could wipe out most, if not all, of their gains.

This has been reaffirmed by more recent research and is obvious when you look at VCT shares through a brokerage platform.

This is something providers don’t hide. In prospectuses, funds will usually include a disclaimer along these lines:

Although the existing Shares have been (and it is anticipated that the New Shares will be) admitted to the Official List and are (or will be) traded on the London Stock Exchange’s market for listed securities, the secondary market for VCT shares is generally illiquid. Therefore, there may not be a liquid market for the Shares, which may be partly attributable to the fact that the initial income tax relief is not available for VCT shares generally bought in the secondary market and because VCT shares usually trade at a discount to their net asset value, the price of the Shares may be volatile and Shareholders may find it difficult to realise their investment. (Source)

Why does this matter?

When investors can only sell at a significant discount, VCT managers are, in practice, unlikely to be punished for poor performance. While I’m indifferent to the fate of the median VCT investor, it’s not only unhealthy for the government to subsidise the distorted allocation of capital, it’s also the exact opposite of what the policy was supposed to do! 

It’s not just liquidity. Even relatively early analyses of the VCT scheme were warning that the focus on tax relief was causing VCTs to underperform standard VC funds. This study pointed to the time constraint of the tax year leading managers to rush the due diligence process and invest in poor opportunities to use up leftover cash. 

Solutions

It was wrong of the government to renew these schemes without assessing their effectiveness. But there is precedent for change. 

EIS’s predecessor, the Business Expansion Scheme, ran from 1983 to 1994. It was replaced because the government believed that it had primarily become a tax avoidance vehicle, rather than a means of supporting businesses. EIS and VCT have gone on the same journey. 

The justification for maintaining the status quo at a time when the UK’s early-stage VC ecosystem is one of the largest in the world feels thin. The original problem these schemes were created to solve increasingly doesn’t exist.

The case for preserving VCT is particularly weak. The scheme itself attracts fewer than 30,000 investors a year and funds’ dismal performance figures suggest it adds little to the world. I would recommend removing all tax relief for VCT investment. These vehicles would likely die and, quite frankly, Britain wouldn’t be the poorer for it.

When it comes to EIS, there’s grounds for partial preservation as part of a reformed system. 

Funds or syndicates that charge anything beyond minimal fees (e.g. legal costs) to investee companies should be barred from the scheme. I would also favour withdrawing the income tax, loss, and IHT relief for all investors, while maintaining the capital gains exemption. This would make the scheme cheaper and drive out many of the bad actors and middlemen. Those left would be better aligned with founders trying to build long-term winners.

It also may be worth revising who is eligible to claim EIS relief. At the moment, the scheme is primarily used by wealthy retail investors who’ve maxed out other tax-free vehicles. While there will be exceptions, I question the extent to which the median lawyer, accountant, or management consultant will ever be a particularly valuable addition to a cap table. The US has a thriving ecosystem of founders and operators who act as angel investors and bring relevant expertise to investee companies. They’re motivated by a desire to make outsized returns, not by the tax system. 

Much of the bad behaviour I’ve described in this piece isn’t confined to EIS and VCT. There’s a longer piece to be written looking at how bad practice, founder unfriendly behaviour, and short-termism remains prevalent in the wider UK ecosystem. Unfortunately, the government, which props up or directly subsidies so much of this bad practice, seems indifferent. It publicly proclaims the UK to be one of the world’s leading venture ecosystems, while giving it the kid gloves treatment you might expect for an emerging sector. 

It’s time for us to decide - are we serious or not?

Subscribe now

Disclaimer: These are my views and my views alone. They aren’t those of my employer (Air Street Capital), my friends, barber, or the four people who follow me on Strava. This is not financial advice. Don’t arrange your finances based on strangers’ Substacks. I have never used (S)EIS (either directly or via a fund), nor have I ever held long or short positions in any VCTs or VCT-backed companies. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned.

Introduction

The cheap coffee, the untouched orange juice, the collection of half-stale pastries, the same two people hijacking the discussion, and the dawning sense that the Earth is rotating on its axis. These are the hallmarks of the dreaded Westminster roundtable. As friends who have routinely suffered through these events, obtaining nothing from the experience, we’ve often struggled to work out what’s in it for their conveners or sponsors. 

We believe that this speaks to a wider malaise in the world of policy. Fuelled by a flood of money from technology companies with vast budgets, we believe that the quality of policy work has been negatively correlated with quantity. As well as being wasteful, it leads to an impoverished standard of debate on issues that matter.

In some group therapy dressed up as a think piece, we canter through the bad types of work that clog up timetables, inboxes, as well as how organisations fritter away their potential influence. And in the spirit of being constructive, we suggest some potential improvements.

While these observations are written with technology policy in mind, we suspect they have broader applicability.

Genres of work

“Convening a conversation”

A favourite of in-house policy teams, this work usually consists of subjecting start-up founders and a loosely defined set of ‘civil society’ representatives to a series of interminable discussions about a set of ‘principles’.

These principles are usually sufficiently vague that it’s basically impossible to object to them without sounding mental (“I hate collaboration”, “innovation sucks”, “AI shouldn’t be developed responsibly”). This renders them practically useless. But the company hopes that by including a constellation of logos on their reports, they can signpost their broad coalition of support when they’re arguing with regulators. 

The problem with this genre of work is that the trade-offs don’t balance out. The ‘research’ is blatantly self-interested, but doesn’t compensate for it by being interesting. Because it’s desperately trying to hide that it’s an advocacy tool, it stops short of surfacing good policy ideas. The only winners are the people running the work, who can signal how busy they are to their bosses.

Rating: 0/5 - no one understands what you do for a living. You are ridiculed in the street and at dinner parties. 

“The government is already doing it … and we agree”

Common among think tanks desperate to appear relevant, this genre involves proposing initiatives or policies that are, in fact, already being implemented by the government. 

The authors begin by identifying an existing government programme or policy, before repackaging it with a few minor tweaks. Look out for vague calls for “expansion”, “enhancement”, “revitalisation”, or, when the authors are really phoning it in, “delivering on current plans”. While this means a government minister will probably speak at your event, it’s a waste of time and may divert attention away from genuine problems. Once you recognise this trope, you’ll see it everywhere.

Rating: 0.5/5 - could have been a tweet. If you want to suck up to ministers professionally, become a SpAd.

“AI will cost 100 million jobs by 2040 and I am sagittarius rising”

You refer to yourself as a ‘thought leader’ on LinkedIn, but you’re just making it up. The ‘implausibly precise future prediction’ is a cousin of the bad economic impact assessment, but it’s even worse. At least the impact assessment is usually trying to answer a theoretically answerable question. 

Typically, this work is produced by either an academic with a book to sell or a new grad at McKinsey with a PowerPoint to make. Because this work provides certainty and specificity, it acts as a frequent source of zombie statistics that clog up corporate presentations and ministerial remarks, years after anyone remembers how they were invented calculated. For example, we’ve seen a Microsoft claim that AI could add £550B to the UK economy by 2035 bandied around various corporate presentations. The source? A ‘study’ where the authors asked GPT-4… #science

Rating: 1/5 - horoscopes for people with PPE degrees.

“We asked people with a vested interest if they like free stuff”

This usually boils down to surveying the potential beneficiaries of a tax change or spending decision. Unsurprisingly, said stakeholders are usually in favour of more tax breaks or a greater share of public spending. In Alex’s recent piece on UK VC policy, this was often one of the government’s preferred ways of measuring the success of its interventions in the venture market. 

The issue is not that this data is useless, it’s just woefully incomplete. Setting aside their self-interest, while industry representatives will know their pain points and can point to domain-specific subtleties that officials miss, they’re often not well-versed in the machinery of government. We have both endured roundtables where this was apparent. 

Industry representatives are also unlikely to consider (or care about) the implications for others of policies that favour them (e.g. funding being diverted). But these studies skip over this and don’t provide any kind of independent economic analysis. 

Unfortunately, this style of policy work is popular because it suits everyone involved. Consultancies and lobby groups can advocate for their clients, while governments like producing documents that say their policies are being received well by other people.

Rating: 1.5/5 - while it might surface the odd useful nugget of information, this work is usually low value and the enemy of good policy-making.

“We’ll bring the raw pie, hope you have a good oven”

The preserve of the well-meaning but naive. They wrongly approach the policy-making process like it’s a seminar and throw out some half-baked ideas or underdeveloped concepts. They then expect policymakers to do the hard work of turning these into actionable ideas through some kind of undefined co-creation process. 

The end product is filled with lofty goals and grand visions, but short on practical details. Think calls for “a complete overhaul of the higher education funding model”, “greater cooperation between the private and public sectors”, or “making it easier to attract the best and brightest to the UK”. These proposals often ignore real-world constraints such as budgets or political feasibility. 

Rating: 1.5/5 - the policy equivalent of “someone should cure cancer”.

“The people we paid say our work is valuable”

The sketchy economic impact assessment is another favourite of consultancies. It typically involves taking either an individual company (that’s usually in difficult regulatory negotiations) or a whole sector and calculating its multi-billion dollar contribution to the economy. 

This usually involves making some … generous assumptions. Tricks include using iffy self-report survey data or choosing the sunniest possible macroeconomic assumptions and working on the basis they’ll never change. The other is simply just to define the ‘sector’ absurdly broadly. For example, lobbyists often claim that the creative industries’ make a £126B contribution to the UK economy every year. They don’t mention that 56% of this number comes from IT and software. 

While the numbers reported are often inflated, the one advantage this genre has is that its authors are forced to share their workings. You can theoretically semi-hijack their model and re-run it using more reasonable assumptions. 

2.5/5 - they’ve given you some of the ingredients to build a spreadsheet. It’s better than a set of principles.

“We gave them a recipe and told them where the ingredients were”

“Implementation” is not simply code for “someone else’s problem”. Imagine you’ve already been given a green light to implement your policy idea. What’s next? What concrete steps do you need to take? Where does the money come from? Which levers of the government do you need to pull? Those who craft their policy recommendations with that in mind are less likely to immediately have them filed under “nice idea, but utterly impractical”.

The end products don’t just say “improve healthcare”. Instead, like the TBI approached their work on the National Data Trust and Digital Health Record, they focus on specific solutions and don’t shy away from getting their hands dirty by developing their technical design. Or take, for example, Reform’s work proposing to establish the Government Data and AI Service, with 19 recommendations on its setup alone.

Rating: 4/5 - answers all those stubborn “how” questions that usually consign policy ideas to the graveyard of good intentions. The challenge lies in delivering this at the right time, given the political, fiscal, and technical realities of the moment. Objectively good, innovative solutions can be ahead of their time.

“We actually established the ground truth on something contentious”

The more time you spend reading policy work or kicking around Westminster, the more apparent it becomes that a lot of it is based on, well, vibes. If enough credentialed people apparently repeat a statement enough, it becomes accepted. This means bad data or arguments from partisans can become accepted as fact, simply because no one has bothered to check the ground truth. 

For example, we see apocalyptic claims made about the prevalence and impact of online misinformation based on comically weak evidence, which have potentially sweeping regulatory ramifications. 

Plugging one of these gaps and using the data as a basis for a set of informed recommendations is potentially impactful. Air Street Capital, Alex’s employer, took this approach towards European universities’ treatment of spinouts. Universities had long claimed that their rapacious approach to founders, forcing them to sign away large equity stakes or extractive royalties agreements to commercialise their work, was exaggerated by critics. 

Air Street surveyed hundreds of founders who had been through the process, obtaining their deal terms and account of the process, showing up the gap between the universities’ claims and the facts on the ground. This also provided an insight into international best practice and suggested a potential model for governments to follow, which the UK government largely accepted following an independent review. 

Rating: 4/5 - particularly effective if you’re campaigning as an outsider, it can make it significantly harder for your opponents to claim you’re a fringe voice or arguing based on anecdote. It does, however, require an ability to select a good question, willingness to invest in sourcing the data, and the patience to accept that it may take a while to get to the top of the agenda. 

“We bothered to check if our brilliant plan is feasible” 

Policy ideas might stand a better chance of making it off the page and into the real world if implementers are consulted first. Groundbreaking.

No matter how intellectually rigorous, not all ideas are usable. At the same time, there is a wealth of insiders, such as former ministers and civil servants, who understand the government’s institutional DNA and can help make policy proposals more actionable. While you need to balance institutional knowledge with challenge, bringing people with you is more likely to result in high-impact follow-up work.

When at Onward, Anastasia saw the effectiveness of this approach first hand. With insiders, including former ministers, on its steering group, the team received their steers on research. This was helping produce practical policy ideas that the government was receptive to. One example is Wired for Success, where her team, among other things, recommended the measures to make the notoriously lengthy Treasury business case procedures more streamlined for certain science and technology projects. These were so lengthy that the business cases would become out-of-date and would have to be resubmitted.

The report caught the eye of Treasury and DSIT ministers, leading to a government-commissioned review on the issue. It was endorsed by the DSIT Permanent Secretary, and all recommendations were accepted.

Rating: 4.5/5 - sometimes insiders can be too wedded to the status quo, shooting down ideas because “that's not how we do things”. But on balance, their reality checks are a net positive. It's the difference between proposing policies for an imaginary government and ones that might actually see the light of day.

Beyond methodology

Policy work often fails because of bad design, timid recommendations, or a lack of realism. But that’s only one half of the picture. The other is moving beyond the piece of research. Too often, companies or think tanks publish a PDF, hold a roundtable, secure a desultory media or newsletter reference, and decide it’s mission accomplished. And the ideas they front up are too generic, too vague, or too inoffensive to matter, because they seek consensus rather than division. This means that even well-designed work accomplishes nothing. Policy isn’t a more secure or better compensated branch of academia. Stacking up publication credits doesn’t come with any rewards. 

There are essentially two models for moving beyond the illusory version of ‘influence’ we often see. At the risk of being crude, you can divide them into ‘insider’ versus ‘outsider’.

The ‘insider’ model boils down to having a forensic knowledge of i) who is actually relevant or wields influence over a certain policy area, ii) understanding their motivations, pressures, and room for manoeuvre, and iii) how to win them over. At its most effective, it leads to individual groups or research organisations being able to shape entire areas of policy. The Tony Blair Institute for Global Change or Labour Together with this government or Onward with its predecessor spring to mind.

The ‘outsider’ approach involves using policy work as a battering ram to shove an issue up the agenda, whether those in the system like it or not. This requires i) a willingness to be confrontational publicly, ii) a thick skin, and iii) an ability to leverage media effectively. As this approach involves a degree of bridge-burning that is culturally alien to most people in policy, it’s less popular. But it can be surprisingly effective. Many institutions (especially in the UK) are so unused to being challenged robustly in public, they don’t really know how to respond to it. 

If your approach can’t be categorised as one or the other, the chances are that it’s not achieving much. This is the reality of most policy work: patchy engagement with peripheral ‘stakeholders’ to build ‘mood music’, matched with cautious external communications designed to avoid offence. 

The thing successful approaches have in common is persistence and discipline. Making progress can take months or even years. Good ‘insiders’ have to navigate their way around a morass of intermediaries, middlemen and associated hangers-on who clog up the corridors of power and will persistently want to have coffee with them. Good ‘outsiders’ need to know which arguments are worth having, when to accept a compromise, and when to push for more.

Disclaimer: These are my views and my views alone. They aren’t those of my employer, my local newsagent, the guy who sells me wine, or anyone else. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned. Any resemblance to real-life roundtable organisers, living or dead, is purely coincidental. Or is it?

Introduction

“British investors targeted in first government advertising push for scale-up boost” blared out the headline of the UK Department for Science, Innovation and Technology press release that landed in my inbox back in March of this year. You can imagine my excitement as I learned that:

“Targeting investors in key financial hubs across the country in cities such as London, Birmingham, Edinburgh, Belfast and Cardiff, the campaign will run on billboards strategically placed 200 metres within investor offices and on key commuter routes. Wrapped copies of The Times will also be delivered to venture capital offices, and digital adverts will run online and across podcasts.”

Seven months later, I remain bitter that my wrapped copy of The Times failed to materialise. This absurdity, however, did not occur in a vacuum. As we see with the current hectoring of pension funds and asset managers to invest in UK equities, attempting to browbeat the financial sector into serving policy goals is currently in vogue. But it’s not new.

From the earliest days of the venture capital industry in the UK, successive governments have attempted to do exactly this. As ‘industrial strategy’ comes back into fashion, I believe the story of the UK government’s patchy efforts to mobilise and shape the industry can teach us a few lessons about incentives, policy design, evaluation, and the distorting effects of lobbying.

Things can only get better?

Until the 1980s, investing in small private companies was a peripheral activity undertaken by a few banks. The Thatcher era brought deregulation, as well as tax relief for early-stage investment. Firms also benefited from the creation of the Unlisted Securities Market in 1980 (the precursor to the AIM), an exchange for small firms with less than three years trading history. The emerging venture boom in the United States likely didn’t hurt either. UK VC firms went from investing £20 million in 1979 to over £1.6 billion in 1989.1

The 1990s in turn saw the birth of a number of schemes that exist to this day, including the Enterprise Investment Scheme and Venture Capital Trusts. The former offered investors in early-stage ventures significant downside protection via tax breaks, while VCTs opened up tax breaks for individual investors in VC funds.

In this era, while the government rolled the pitch in the industry’s favour, it stayed on the sidelines. 

This all changed with the return of the Labour Party to government in 1997. In its first term, the Blair Government created a series of vehicles to channel hundreds of millions, and later billions, of pounds of public money into early-stage businesses. 

The government was motivated by a desire to close the ‘equity gap’ - the situation where the capital needs of some early-stage businesses are too high for angels, too risky for banks, but too small to attract VC firms. 

Originally a complaint dating back to the Great Depression, it received fresh urgency when the Bank of England in 1996 and the Confederation of British Industries in 1997 published reports warning that the growth of the VC industry in the UK had done little to help emerging tech firms. A common criticism was that the industry disproportionately focused on later-stage, lower risk activity like management buy-outs, rather than supporting innovators.

A 2003 Treasury Report argued that this gap was severe for businesses seeking up to £2 million and warned that: “This equity gap is a barrier to productivity growth, as it can stifle the development of innovative start-up and early-stage businesses, and can constrain the supply of capital for some established businesses that are seeking to modernise or diversify their activities.”

Outlined in a 1998 white paper on competitiveness, the new Labour Government spun up a number of initiatives, designed to use government money to crowd in private investment. These included:

• UK High Technology Fund - a £126 million fund of funds, catalysed by a £20 million government commitment, which invested in nine funds.

• Regional Venture Capital Funds - £226.5 million invested over nine funds, designed to showcase the potential for private sector investors to achieve a commercial rate of return outside London, featuring £74.4 million of public money.

• Bridges Funds - £40 million fund for investment in the 25% most disadvantaged areas of England, half of which was provided by the government.

• Early Growth Funds - £91 million, catalysed by a £26.5M government commitment, to demonstrate the potential of investing in early growth businesses. 

There was an attempt

These efforts did not succeed. While the government remained cagey about performance, a 2009 National Audit Office review was scathing.

It found that the UK High Technology Fund’s interim internal rate of return was -9.7% and the RCVFs’ -15.7%, while the Bridges Fund had performed a little better at +7.7%. Bridges, however, had far fewer constraints on its investment activities and (bizarrely) was able to invest in property, which became the primary driver of its returns. 

To lure in private sector investors, the government had agreed to bear the brunt of any losses and to offer them preferential treatment in the event of any returns. This meant that the original £74 million public investment in the RCVFs was worth £5.9 million - a -93% return on investment.

All the while, the department was paying out fees to professional fund managers, which ate up 36% of the money invested in the RCVFs, 17% of the High Technology Fund, and 29% of the total invested in the Bridges Fund.

Even before the NAO report, there were red lights on the dashboard. In 2007, when the RCVFs had been up and running for six years, half the capital remained uninvested and the external managers were complaining about a lack of suitable management teams in some regions. 

The manager in the north-east said that “it has been tough getting the money out there”, while in the East Midlands warned that “there is no lack of capital, but there is a lack of good management teams to invest in”. The government’s defence was that “some 283 SMEs have been supported so far, indicating an ongoing requirement for equity investment of [this] type”. 

This is a wrong-headed argument that remains a staple of policymaking to this day. 

Firstly, the act of spending money is a bad metric for … somewhat obvious reasons.

Secondly, if these teams or businesses were genuinely good, there is a chance that the government may have just succeeded in crowding out private investors. 

In fact, research was already appearing in Europe from the mid-noughties pointing to how government investors that were simply content to receive their money back were crowding out more disciplined private investors. One study of fourteen European countries found that every additional dollar from government funds led to a dollar less from their private counterparts.2

But was government content just to receive its money back? 

Well, we just don’t know. As the NAO put it: “no clear financial objective was set for the impact of the funds to the taxpayer, such as whether they were expected to break-even and over what timescale, and the Department did not specify objectives for wider economic benefits apart from the Bridges Funds” and noted that they had made no attempt to evaluate the funds in eight years.

This was backed up by one RVCF manager at the time, who said in 2006: "I wonder whether the publicly backed funds aren't doomed to fail. On the one hand they are supposed to be about fostering entrepreneurship but on the other they are supposed to make money. In some respects these are conflicting objectives. Certainly they pull in different directions.''3

Mr Anonymous Fund Manager points to a central tension in UK government policy on venture capital - the lack of clarity around objectives. 

Is the purpose of this support to back strategically significant proto-winners or to act as a small business financing scheme?

The real equity gap was the friends we made along the way

Based on its heavy use of narrow regional constraints and choice of evaluation metric, it appears to be the latter. The other real ‘tell’ is that when you look at the big technology strategy documents from the government in this era, this work is simply not mentioned in them, beyond the odd passing comment. If the purpose of this work was to back strategically important companies, then there’s no written evidence of it.

This makes venture funding almost certainly the wrong vehicle. Discussion of the equity gap often focuses on the inability of ‘viable’ businesses to get money - in fact a current civil servant used exactly this turn of phrase when we were debating this on X.

But ‘viable’ means different things to the management of a small business and a partner at a VC firm. VC is not suitable for most enterprises: unless a business has the potential to be a runaway success story, investors simply can’t justify the risk of backing them. Turning a profit and becoming a respectable regional employer is great, it’s just a different thing! Confusing these definitions of viability results in attempts to leverage a type of financing for a purpose it was never intended for.

The evidence base that there was this ‘equity gap’ that needed plugging by the government was, at the very least, contentious. 

When these policies were first devised, there was no comprehensive database of VC transactions in the UK. Rather than asking investors for data, the government heavily relied on asking businesses if they were able to raise all the money they needed from investors. 

In 2006, now-defunct research firm Library House created a transaction dataset for the first time. They found that there was no shortage of deals in the £250k to £2 million range and that “there is no ‘gap’ in the range of funding deal sizes available to growing companies in the UK”.

As they put it: “A comparison of the funding requirements of companies with subsequent funding received shows that a significant gap remains between desire for funding and actual investment. However, this phenomenon is only partially related to the level of funding available and is more reflective of the fact that the majority of companies seeking funding simply do not have the potential required to warrant investment by an investor motivated by financial gain.”4

Yet, they persisted

Did the government learn from these early mistakes? Yes and no.

Even before the NAO report dropped in 2009, some of the government’s work was trending in a more promising direction. From 2006 onwards, the government began acting as an anchor investor in new VC funds via the Enterprise Capital Fund programme. This has helped bring around 30 new fund managers into business and its performance has fared significantly better than its predecessor initiatives. 

The British Business Bank, which took over the management of this work after its creation by the Coalition Government, commissioned a third party analysis of the ECF scheme in 2021. It concluded that each £1 of cost brought £2.62 in net economic benefit. I have my gripes with the methodology employed, which largely consists of stringing together a bunch of correlations, but we can see from the accounts in the BBB’s annual reports that the scheme is definitely making money. 

There is evidence of the scheme helping to bring in genuinely interesting new managers into the business. Its diligence process is thorough, and the BBB draws in private capital by taking less than its share of carry (although it gains downside protection via preferential returns).

However, an unfortunate proportion of this work has gone towards subsiding incumbents who should be able to raise by themselves, or in some cases, had successfully raised funds before the creation of the ECF programme. 

The BBB has also ignored the NAO’s call for greater transparency and releases no data on how individual funds have performed. Privately, I’ve heard stories of funds that have literally never returned a penny to investors continuing to receive BBB backing “because they work on deep tech and therefore it’s understandable that they don’t have any DPI yet”. This is expensive SME subsidy reincarnated.

In other areas, there’s more evidence that old habits continue to die hard. Despite its separation from Whitehall, the BBB has not been able to resist the age-old temptation of creating yet more regionally-focused equity financing vehicles, managed by outside fund managers. This is despite the NAO warning against regionally-constrained funds and common-sense logic suggesting a narrow regional focus is bad for diversification. 

It’s not just the BBB. An alphabet soup of government bodies clubbed together to create yet another early-stage innovation fund (run by external managers). Troublingly, it points to job creation as one of its metrics of success.

As the UK’s VC industry matured and the early-stage ecosystem exploded, the original ‘equity gap’ thesis became less and less plausible. Instead of being sent to live on a farm (no, you can’t visit it), it mutated. While the ECF programme still employs the ‘equity gap’ as part of its rationale, it has been superseded in the wider discourse by the ‘funding gap’. This is an alleged shortage of growth capital for UK companies.

The founding charter of the funding gap is the government’s 2017 Patient Capital Review, the product of much industry lobbying. In this era, investors were upset that promising UK companies often moved to the US when they hit growth-stage, while less promising ones struggled to raise. Instead of judging this as the product of i) some companies being less good than their early investors liked to admit ii) growing businesses consciously choosing US partners who could provide better market access, this was (much like the original equity gap) presumed to be the product of market failure.

The Patient Capital Review, much like earlier work in this genre, eschews any kind of independent economic analysis in favour of consulting a collection of founders, asset managers (including one Neil Woodford), and VCs with vested interests. In short, the government asked a group of people if they wanted the government to mark up the value of their assets - and you’ll never guess what they said! 

This led to proposals for a proposed government vehicle that could invest ‘patient capital’, as part of a transformed landscape:

British Patient Capital was duly set up by the BBB, with £2.5 billion to invest in venture funds or directly into promising companies reaching the growth stage. BPC also has the £425 million Future Fund: Breakthrough initiative to invest directly in growth-stage companies.

Catalysing investment… or distorting it?

Thanks to the Cameron and then Sunak-era embrace of technology as a source of economic growth, this work rhetorically has more of an R&D flavour than the SME jungle of New Labour-era initiatives, but there’s still a sense of directionlessness under the hood. 

Considering the alleged purpose of this work is to plug the ‘funding gap’ and deliver commercial returns (no potential conflict acknowledged…)  it’s worth questioning why government cheques are being thrown into multibillion dollar incumbents that quite obviously do not need the money or why it is anchoring some firms’ fifth funds. In the former case, there is no funding gap. In the latter, it’s worth questioning if a fifth fund that needs government support is likely to achieve commercial returns. 

It’s also worth questioning why the BPC is supporting region-limited university spinout funds, which can hardly be judged to be late-stage capital. They also, again, underscore government’s inability to understand the damage regional constraints pose to a venture model where diversification is critical.

However, The Future Fund: Breakthrough (FFB) is by far the most egregious example of the BBB defying both the logic of the market and common sense. The idea was simple. By acting as an investor in a funding round for a company at growth-stage, the government would give private VCs the confidence to invest their own money, while the final funding round would be bigger. This would help British companies working on crucial R&D to scale, without having to move to the US.

While this might sound logical, it in fact describes the perfect reverse selection mechanism. A scheme designed to bolster companies unable to find investors with conviction is disproportionately likely to select for those with weaker growth prospects. Government is simply creating a false demand signal, preventing the market doing its thing, and talent and capital being recycled in more profitable directions. 

The only way this logic doesn’t hold is if you believe that the BBB has a significantly better ability to identify potential winners than private firms - something that even the organisation itself has not claimed.

The BBB’s own evaluation of the scheme concluded that it was very successful. But the document is littered with red flags.

Firstly, it uses a series of input measures. The fact that it exceeded its goal of investing alongside £500m could be interpreted as a sign of success or simply a reflection of the glut of private capital on offer at the time. My guess the latter is the case, which does, again, undermine the rationale for the entire programme.

At times, it comes incredibly close to spotting the adverse selection, such as its celebration of essentially bailing out at least one collapsing business. 

Considering the volume of private capital available in the UK now, when the government is providing 20% of a company’s Series B funding round, as the BBB celebrates, it’s either a sign that the business is bad or that private investors are being crowded out.

The final cherry on top is the inclusion of a put option, which gives the BBB to ability to cash out if the business becomes insufficiently UK-focused. This is highly unconventional, gives a potentially small investor effective veto rights over corporate decision-making, and seems … political. It also signals that the priority for this work is a narrow conception of ‘UK plc’, rather than real global leadership. The new ECF criteria for investment have also beefed up their UK requirements - a retrograde step.

This only scratches the surface - there are so many funds, and vehicles that it is difficult to keep track and evaluate their performance comprehensively. These include a government UK Innovation Fund (run by external managers) that launched after receiving only £180 million of the planned £1 billion in private commitments and drove good returns after allocating only 40% of its portfolio to UK companies. There’s also a UKRI vehicle that seemingly randomly gave £42 million in non-dilutive grants to a single nanotech company in Northern Ireland, and yet another BBB Managed Funds programme. 

When you look at this work in the round - it’s just unclear what it adds up to. It still doesn’t act as an effective small business finance mechanism, while propping up ropey Series B firms doesn’t scream ‘science and tech superpower’.

Reflections

It’s easy to conclude that much of the above is ancient history and that we’re all a lot more sophisticated, now the industry and the ecosystem have evolved. But as the government takes a more interventionist turn, there are a few crucial lessons that extend beyond the venture market.

For a start, we need a significantly better approach to setting goals and evaluating progress. “Closing the equity gap” is a means not an end. Government needs to have clear and consistent reasons for supporting enterprises that are otherwise unable to obtain funding, beyond automatically asserting market failure.

A lack of real objectives leads to policy evaluation by input measurement (“did we spend money?”), which is rarely helpful. Similarly, polling the recipients of said money should not be a measure of success. 

Ultimately, we need to clarify the central tension between backing winners versus subsidising the rest. If the real priority is the latter, then we should probably just junk all of this work, channel this money into bulking up the government’s existing SME loan guarantee schemes, and stop paying fees to private managers for the vibes. It might be less ambitious, but it would be significantly cheaper.

If the goal is to back proto, venture-scale winners, there needs to be a transparent way of evaluating whether we’re actually doing this and a policy of cutting off losers. 

The public should be able to see returns and the BBB shouldn’t be able to hide behind multiplier figures that consultants they’ve paid have calculated. After a maximum of two funds, ECFs should be able to show signs of success, after which they should be able to raise without the BBB anchoring them. If a start-up isn’t able to raise a Series B, maybe this isn’t evidence of a ‘funding gap’ and instead a sign that they’ve failed.

This also means that the government will have to stop outsourcing policy-making to self-interested market participants. Policymakers often act as if small businesses, universities, and British investors are neutral arbiters on policy questions (unlike evil US tech companies). This is driven by a long-standing and misplaced Westminster awe at people who work in finance, a lack of inhouse expertise, and a wrong-headed belief that small businesses or universities won’t lobby as aggressively for their interests as everyone else.

Finally, we need to avoid the trap of assessing each element individually and losing sight of the macro.

The original rationale for government interventions that now date back 23 years was to crowd in more private capital and to close the ‘equity gap’. The BBB today still talks about crowding in private capital and closing the ‘equity gap’. 

Given the passage of time and the money expended so far, it seems reasonable to ask if the problem still does exist and, if so, if the current policy toolkit is the right one. 

I’m inclined to argue that the government is operating on outdated assumptions and that it should focus its efforts elsewhere. There’s no shortage of real problems, including poor rates of business investment, low competitiveness levels, bad infrastructure, and terrible government procurement. Fixing these will require more creativity than smashing the subsidy button. Industrial strategy in the UK has historically failed to deviate from this approach. Let’s see if this time really is different.

Disclaimer: These are my views and my views alone. They aren’t those of my employer (Air Street Capital), my friends, my friends’ tennis partners, or my family. I’m not an expert in anything, I get a lot of things wrong, and change my mind. Don’t say you weren’t warned. I’m also not saying that everyone who has received money from the government is bad or undeserving etc. etc. to pre-empt any overly-personal responses.

Photo: a ‘bridge to nowhere’ near Castrop-Rauxel, Germany. (Der Hessi/German Wikipedia)