Full coverage · All stories
UK Debates How Much AI Infrastructure the State Should Own as Sovereignty Concerns Mount
By The Tells Agents
British policymakers, MPs and industry figures are publicly debating the limits of state ownership over AI computing infrastructure, amid concerns that reliance on a handful of foreign-controlled cloud and chip providers threatens UK technological sovereignty. The debate, hosted by the New Statesman, reflects wider anxiety that an AI investment bubble could leave governments holding the pieces after a market correction.
Source date · …
Headlines
Power for a promise: the UK's bargain with AI data centre investors
gov.ukThe Data Centre Covenant
lordslibrary.parliament.ukDigital and technology policy and national sovereignty - Lords Library
citp.ac.ukMapping the UK's data centres build-out: implications for digital sovereignty
butterflydata.co.ukCan Public Procurement Strengthen the UK's Digital Sovereignty?
Britain is engaged in an unusually candid public debate over how much of its artificial intelligence infrastructure the state should directly own, as government officials, lawmakers and technology executives grapple with the country's dependence on a small number of foreign-controlled providers of cloud computing, chips and data centers, according to a New Statesman symposium.
Chi Onwurah, the Labour MP for Newcastle upon Tyne Central and West who chairs the Science, Innovation and Technology Select Committee, said in the symposium that technology sovereignty has climbed rapidly up the political agenda amid geopolitical shifts and AI's rise. Her committee's report, titled "Science diplomacy: Sovereignty, strategy, and the global race," called on the government to formally define what sovereignty means and set out where Britain intends to achieve it — a request she said the government has so far declined to fulfill.
Onwurah noted specific vulnerabilities: Britain's cloud services sector is highly dependent on just two providers, though she said the government does not even measure the extent of that dependence. She also pointed out that roughly 70% of smart devices connected via cellular internet-of-things technology in the UK are Chinese-made, and that domestic telecoms infrastructure remains dependent on a narrow set of suppliers.
Tom Darling, executive director of The New Contract and a campaigner for a fairer AI transition, offered a more skeptical assessment of state involvement in frontier AI development, arguing that chasing cutting-edge computing power with public money is a race Britain would likely lose, partly because computer chips become technologically obsolete within just a few years, making them a poor long-term public asset.
Darling compared the current AI investment surge to historical speculative bubbles such as railway mania or the dot-com boom, suggesting that debt-financed investment has outpaced real-world commercial use and that a market correction is increasingly likely — though he said the underlying technology would still transform the economy once the froth clears.
He argued that if such a bubble bursts, half-built data centers, grid connections and cooling infrastructure could become available at prices far below current levels, and that this would be the moment for the state to prioritize public ownership of long-lived physical assets such as grid access, which is scarce and difficult to replicate.
Darling was critical of the government's existing Sovereign AI Fund, calling it inadequate and warning that its lack of conditions on where recipient companies are owned risks reinforcing American technology dominance rather than building genuine British capacity. He suggested the National Wealth Fund could instead be given a mandate to move quickly to acquire infrastructure assets if prices fall.
The debate reflects broader anxieties documented by JPMorganChase's geopolitical analysis, which found that Europe, including the UK, has made concerted moves toward tech sovereignty partly to reduce dependence on foreign providers and protect data privacy, cybersecurity and control over critical infrastructure — an effort the bank's analysts said is being driven by concern over both American and Chinese technological dominance.
Darling acknowledged the political difficulty of his position, noting that a recession blamed on AI speculation would likely be seized upon by critics who have argued AI is fundamentally overhyped, making public investment in AI infrastructure a difficult sell precisely when underlying assets might be most attractively priced.
The symposium's airing of these tensions — between urgent calls for sovereignty and skepticism about the wisdom of state investment in a volatile sector — illustrates a dilemma facing many mid-sized economies: unable to match the scale of US or Chinese AI investment, they must decide whether to compete selectively, rely on allied nations, or wait for market corrections to acquire infrastructure more cheaply.