Stargate was to be the world’s largest AI investment: a $500 billion infrastructure Project “Securing US leadership in AI”. Never shy of hyperbole, its lead backer, ChatGPIT-maker OpenAI, promised “massive economic benefits for the entire world” with features to help people “use AI to elevate humanity.”
Now, OpenAI Appears Opting out of one part of the deal – the expansion of a major datacenter spanning a large swath of land in Abilene, Texas, has become one of the most visible manifestations of the frenzy of investment in the chips and power plants needed to build and operate AI. Negotiations over project financing have hit a wall, as has a timeline for when the expanded capacity could come online.
This may be fine for OpenAI; It may possibly find another datacenter. That’s less okay for Oracle, OpenAI’s partner on the project, which has already spent billions on hardware for the site. This is one of several cracks appearing in the capital side of the AI economy that are unsettling investors.
both companies to pass Said The development won’t derail their AI plans. He also said that a month ago, when a separate $100 billion deal was struck between OpenAI and Nvidia, the world’s largest maker of chips that trains AI models and answers the billions of questions people ask daily.
The future of such deals is becoming increasingly important for the global economy. According to Bloomberg, future datacenter leases agreed by the largest cloud computing companies (including Amazon, Oracle, and Microsoft) have increased by nearly 340% in two years and now top $700 billion. That’s a lot of money if the technology doesn’t begin to deliver on its promise of increasing economic productivity. On Friday, more than three years after the launch of ChatGPT that sparked AI hype, the UK reported zero GDP growth for January.
On Monday, the Guardian exposed another crack in the AI edifice. Major UK AI deals, many of which were announced with great fanfare during Donald Trump’s state visit last September, are not as described in government and corporate press releases, an investigation has found. Major projects are delayed or impossible, significant “investments” are actually vague agreements between mostly US tech companies, being touted by ministers as an engine for economic growth.
If the cracks in this datacentre boom grow, the consequences will be that the UK will end up without the AI infrastructure needed to survive in the global economy and the more serious risk will be that the entire AI bubble bursts in a repeat of the 2001 dotcom crash which could shock the world economy.
“There has been a lot of blind optimism around building AI infrastructure,” said Andy Lawrence, executive director of research at the Uptime Institute, which inspects and evaluates datacenters. “While there is an incredible boom underway, with construction taking place on a scale never seen before – it has also long been clear that many projects will either not go ahead, or will take much longer to be built and begin operation than claims suggest. Because of the high stakes and high rewards in AI, it has attracted speculators who promise investment but have little experience in the field.”
Most symbolically, the Guardian’s investigation showed a government site in Loughton, Essex Said Will host “the largest UK sovereign AI datacentre” by the end of 2026. The then Technology Secretary, Peter Kyle, called it “a new start for our economy and working people”. A year later it was still being used as a staging yard and the chances of it being open once the bills were paid were almost zero. Following an investigation by the Guardian, Enscale confirmed it had purchased the land on which the computer is to be built – eight months later said it happened In January 2025. It still does not have planning permission, but said on Friday it was planning to start construction before July and switch on the datacentre between April and July 2027.
The AI deals come amid growing ties between US tech corporations and senior politicians in the US and UK. Donald Trump’s top AI advisers include David Sachs and Sriram Krishnan, both of whom have recent histories as tech investors. In London, OpenAI hired former Chancellor George Osborne; Anthropic and Microsoft hired former Prime Minister Rishi Sunak; Peter Mandelson was the owner of a consultancy that lobbied for Palantir; And the Tony Blair Institute has received funding from the foundation of Larry Ellison, the billionaire owner of Oracle.
These figures have helped create an AI policy in which the UK has essentially agreed to become a platform for US-designed hardware that is mostly being rented to US tech companies. The UK government says it is building “sovereign AI infrastructure”, a controversial definition ranging from UK-owned hardware and data so it can retain control over a piece of critical national infrastructure in a world of unstable international alliances, to AI Minister Kanishk Narayan’s more flexible definition as “strategic leverage” so the UK can “ensure continued access to critical inputs”.
In the UK this means being dependent on the US. As Nvidia Chief Executive Jensen Huang said during Trump’s state visit last September: “The United States must lead the entire AI technology stack.”
Former Deputy Prime Minister Nick Clegg put it more bluntly that week, calling Britain a “technically vassal state”. Clegg this week became a board director at Enscale, the UK company involved in the Lawton AI deal, where its client is Microsoft – part of the US tech hegemony whose power he lamented six months ago.
On BBC Radio 4’s Today program this week, Enscale’s senior vice president, Imran Shafi, was asked whether its Essex datacenter would be live by “Q4 2026” as promised. He replied: “The time it will go live will be the time we have approved with our customer.”
Meanwhile, Narayan defended the broader pace of progress. “What we are saying is that we are making solid progress,” the minister said. told CityAM. “We already have live datacentres in Lanarkshire. We have hoes in the ground in parts of the north-east.”
Narayan might consider the example of the current recession in Texas. Billions of dollars were promised, construction began, billions of dollars worth of equipment was purchased, and then OpenAI walked out, leaving its partners in the uncomfortable position of finding another giant AI company to work with.
OpenAI, it was informedA new chip model was wanted: and by the time manufacturing in Texas was completed, the hardware purchased by Oracle would no longer be state-of-the-art. It was like buying a lot of iPhones right before a far more powerful model was launched. The speed at which chips are aging is further undermining the UK government’s claims of massive AI investment. This is describing in cash terms the “investment” which is mostly computer chips. Chips are not money – they depreciate, possibly even and faster Most tech companies say they will.
This means it matters when the datacenter in Essex or the AI hub in Lanarkshire should come online. By the time they are ready and the source of additional power is ready, will the leap in AI system design mean that running 2025 chips will be like owning a propeller plane in the jet age? Or if the deals announced are related to future chips, will they be available? Iranian drone attacks have already affected supplies of helium from Qatar, which chipmakers need. What will happen if China disrupts supplies from Taiwan?
“Datacenters, especially large high-density AI, are very complex engineering projects,” said Lawrence at the Uptime Institute. “Few are completed in less than two years, and usually it takes much longer. It is not unusual for some projects to be delayed for years or postponed indefinitely.”
The final component here is banks. There are chips from Enscale, and chips from other datacenter companies. took advantage of. These operators have secured billions of dollars in loans based on their graphics processing units (GPUs). At least in Enscale’s case, this loan will go to finance its UK buildout, but when will that loan come due? If it can’t be paid, what will happen to Enscale or the financial institutions potentially involved in finding buyers for the old chips?
A spokesperson for Enscale said it “works with established financial counterparties and maintains disciplined governance around financing decisions. We take a conservative approach to building long-term infrastructure for our financing.”
Alvin Nguyen, an analyst at Forrester, said: “People who are lending money, financial institutions, they are taking on a lot more risk because the chips have a lifetime.”
The boom in datacenter investment represents one of the biggest infrastructure gambles of this or any era. Whether the scaffolding yard in Loughton becomes a real AI factory could tell us a lot about who will win and who will lose.