|
In this Commentary
This Commentary argues that Britain’s £500 million Sovereign AI Fund is a bold and necessary strategic step, but warns that public capital in a small, networked AI ecosystem must be governed transparently. Without open competition and robust safeguards, industrial policy risks entrenching insider power rather than national prosperity.
A Small Ecosystem with Large Consequences
Britain’s AI sector remains relatively young. It is sophisticated, energetic and increasingly global, but it is still compact enough that many of the principal actors know one another. Founders know investors. Investors know advisers. Advisers know ministers. Academics sit on boards. Civil servants rotate through policy circles populated by the same people who later advise funds or companies. Conferences, labs, committees and private dinners form a recognisable circuit.
This is normal in an emerging industry where expertise is scarce and experience is concentrated. Every new sector begins with tight networks. Talent clusters. Trust networks form. Relationships matter.
Yet because this is normal, governance becomes essential.
When everyone knows everyone, decisions made in good faith can still look partial. Companies selected on merit can appear pre-selected. Legitimate judgments can lose public confidence if the process that produced them is opaque. Legitimacy can evaporate in the absence of transparency regardless of whether wrongdoing has occurred.
That distinction matters.
The issue is not whether any company deserves support. Some almost certainly do. Nor is it a suggestion that specific individuals could act improperly. The deeper issue is whether sovereign capital is being allocated through institutions strong enough to resist the gravitational pull of proximity, familiarity and status.
Public money cannot rely on private assurances.
Why Procedure Is Substance
There is a recurring temptation in British policymaking to dismiss procedural questions as secondary. We are told to focus on outcomes, not process. If the right companies are funded, why worry about the mechanics?
Because in strategic markets, process is substance.
The method by which decisions are made determines who gets seen, who gets heard, who gets introduced, who receives the benefit of doubt and who never enters the room. Informal systems reward those already embedded within them. They privilege fluency in elite codes over raw capability. They select for social access as much as technical merit.
Once that pattern hardens, it reproduces itself.
The firms chosen in the first round become the firms everyone assumes are the leaders. They attract more private capital, better recruits, greater media attention and easier access to government contracts. Their early endorsement compounds into market advantage. Meanwhile, equally capable challengers struggle to be noticed.
This is how concentration often begins: not through explicit favouritism, but through seemingly reasonable choices repeated inside narrow circles.
If Britain wants an AI economy defined by competition and invention, it must pay close attention to the architecture of selection.
Three Rules That Should Be Non-Negotiable
The Sovereign AI Fund should therefore operate under principles clear enough to command confidence and robust enough to survive scrutiny.
Transparent Standards
Government must state plainly what it is trying to back.
Is the aim frontier model development? Commercial traction? Public-sector utility? Strategic autonomy? Regional regeneration? Export potential? Scientific spillovers? Defence relevance? Productivity gains in critical industries?
These goals are not identical. A company optimised for cutting-edge research may look very different from one built to transform NHS workflows or modernise manufacturing supply chains. If ministers and fund managers do not specify the weighting of criteria, outsiders will naturally suspect that criteria were created after decisions had been made.
Clear frameworks protect everyone: applicants, taxpayers and those selected.
Credible Safeguards
In a close-knit sector, relationships are unavoidable. That is why declarations of interest, recusals, external reviewers and independently documented decisions are not bureaucratic extras but the minimum price of legitimacy.
Where conflicts are real, they must be managed. Where they are perceived, they must be explained. Silence invites cynicism. Disclosure builds trust.
Britain has enough talent to do this properly. It should do so visibly.
Open Contestability
Sovereign funds must never become concierge services for the connected.
Britain’s next strategic champion may not sit in the obvious postcode. It may not be backed by fashionable funds. It may emerge from a university spinout outside the Golden Triangle, a specialist enterprise software team in the Midlands, a defence-adjacent start-up in the Northeast, or a technical founder ignored by current market fashions.
If access depends on being known in advance, Britain will miss the people such a fund was created to find.
The Economic Cost of Insider Allocation
The danger here is not just moral or political. It is economic.
When capital repeatedly circulates through the same social graph, markets become less intelligent. Novel approaches are screened out before they are tested. Unconventional founders are underfunded. Incremental bets crowd out bold ones. Status substitutes for evidence. Reputation substitutes for results.
Britain knows this story in other sectors. We have often mistaken polish for competence and familiarity for excellence. We should not repeat that error in AI, where the frontier is moving quickly and breakthroughs may come from unexpected quarters.
The cost of getting this wrong would be high because AI markets are path dependent. Early financing decisions can determine who accumulates data, who recruits scarce talent, who secures enterprise customers and who gains the compute resources necessary to improve products. Initial advantages compound fast.
In such an environment, poor allocation in year one can distort competition for a decade.
Britain’s Strategic Choice
The wider geopolitical context makes this more urgent.
Across the world, nations are recognising that AI is not just another sector. It is foundational infrastructure. The countries that shape it will influence standards, security, industrial competitiveness and the future distribution of wealth.
The United States has significant advantages: deep capital markets, hyperscale cloud providers, elite universities and a culture that tolerates outsized risk. China has pursued a more state-directed path, combining industrial strategy, infrastructure investment, strategic finance and determined cultivation of national champions.
Each model has strengths and weaknesses. But both understand a central truth: technological capacity at this level is too important to leave unattended.
Britain cannot replicate either model wholesale, nor should it try. Our task is different. We need a distinctly British approach that combines strategic intervention with open competition, strong institutions with entrepreneurial energy, public purpose with private dynamism.
That is a harder balance to strike. But it is the right one.
Varoufakis and the Warning from Techno-feudalism
Yanis Varoufakis has argued in Technofeudalism that contemporary capitalism is mutating into something closer to a feudal order. In his account, markets are increasingly hollowed out by digital gatekeepers who control platforms, data flows, infrastructure and attention. Economic life no longer revolves primarily around competitive production, but around rents extracted by those who own the digital estates on which everyone else depends.
One need not accept every element of the thesis to recognise the force of the warning.
Power in the digital economy does tend to concentrate. Network effects are real. Compute access is uneven. Distribution channels are dominated by a handful of firms. Data advantages can be self-reinforcing. Once scale is reached, incumbents become difficult to dislodge.
If Britain’s sovereign strategy just channels public legitimacy toward already privileged networks without broadening competition, we risk reproducing this pattern domestically. We would socialise prestige while privatising upside.
That would be a mistake.
What Success Would Actually Look Like
A successful Sovereign AI Fund would be judged not by headlines on launch day, but by structural outcomes five years from now.
It would have backed companies across the country, beyond the usual enclaves. It would have supported the full breadth of the stack: applications, infrastructure, specialist models, developer tools, cybersecurity, health technology, defence systems and productivity software. Done well, it would have mobilised private capital rather than substituting for it, improved public services through genuine deployment rather than perpetual pilots, and helped build British firms able to compete globally while remaining anchored at home.
Most importantly, it would have increased the number of serious contenders.
That is what effective industrial policy should do: widen the field, create more credible winners than the market would have produced on its own, and deepen national capability rather than narrowing opportunity.
By contrast, failure would look different. A small circle repeatedly favoured. Opaque rationale. Weak additionality. Companies selected because they were already visible. Limited regional spread. Sparse downstream impact. A fund remembered as political theatre rather than national strategy.
Britain cannot afford the latter.
A Better National Instinct
There is often a curious British hesitation around backing our own capabilities. We celebrate invention but distrust scale. We admire entrepreneurs until they become powerful. We speak of strategy but recoil when strategy requires choices.
The Sovereign AI Fund suggests that instinct may finally be changing.
That is welcome. A mature nation should be willing to invest in sectors central to its future. It should be willing to shape markets where strategic dependence would otherwise grow. It should understand that neutrality is sometimes just passivity dressed up as principle.
But strategic confidence must be matched by institutional seriousness.
If government wants public trust for activist economic policy, it must show that activism is disciplined, fair and accountable. Otherwise, every intervention becomes vulnerable to the charge that it is simply patronage with modern branding.
Takeaways: The Castle Walls Must Stay Open
Britain should celebrate the ambition behind this fund. It represents a recognition that AI will help determine economic power in the decades ahead and that the state cannot remain a spectator.
Yet ambition without integrity quickly curdles. A sovereign fund without transparent standards, visible safeguards and open access would not strengthen capitalism, but erode confidence in it. It would teach talented outsiders that the game is closed and confirm the suspicion that in modern Britain the future is often brokered privately before it is announced publicly.
That outcome is avoidable.
We can build an AI strategy that is competitive rather than clubby, national rather than captured, bold rather than performative. We can use sovereign capital to widen opportunity, accelerate adoption and create real domestic strength.
But only if the rules are as serious as the rhetoric.
If Britain gets the Sovereign AI Fund right, it could help shape a more open, innovative and resilient technological economy. If it gets it wrong, Varoufakis’s warning may look less like theory and more like diagnosis: a new techno-feudal order in which power concentrates, access is rationed, and the future belongs chiefly to those already inside the castle walls.
|
13 hours ago
Facebook
Twitter
Google+
LinkedIn
