The UK government's ambitious plan to 'mainline AI into the veins of the economy' has been met with enthusiasm, but a closer look at the investments reveals a more complex picture. As reported by The Guardian, the drive is riddled with 'phantom investments' and shaky accounting, raising questions about the true value of these investments.
At the heart of the government's AI plans are two companies backed by the $4tn tech giant Nvidia: London-based Nscale and the US company CoreWeave. Since 2024, successive Conservative and Labour governments have proclaimed massive deals to build new datacentres, create thousands of jobs, and construct a supercomputer. However, a Guardian investigation has found that the money isn't necessarily real, the datacentres may not be new, and the jobs are unaccounted for.
The UK's Department for Science, Innovation and Technology declined to answer detailed questions but said it 'rejected these assertions.' A statement added that the AI sector has attracted more than £100bn in private investment since the government took office, with the AI sector growing 23 times faster than the wider economy last year. However, the department also acknowledged limitations to its oversight, stating that it was 'not playing an active role in auditing these commitments.'
Case Studies: CoreWeave and Nscale
CoreWeave's £1bn investment in the UK was hailed as a major deal, with the company announcing that it would bring 'two new datacentres to our shores.' However, planning records indicate that CoreWeave built no new datacentres at the announced locations. Instead, the company became a customer of two existing datacentres, one built in 2002 and one built in 2015, and deployed Nvidia chips that it had paid for.
Nscale's flagship project, a supercomputer site on the outskirts of Loughton in east London, was announced in January 2025. The company said it had 'signed a contract' to complete the datacentre by 2026, but the proposed site was still being used as a scaffolding yard when the Guardian visited in February. Nscale submitted a planning application to build there at the end of February, after the Guardian had begun to make inquiries.
Cecilia Rikap, a professor of economics at University College London, said that 'big tech companies artificially inflate datacentres' job creation and economic impact to please governments like the British one, which are desperate to claim they are making the economy grow.' She added that 'these are phantom investments,' and that similar things are happening around the world.
Expert Analysis
Dr Kat Jones, director of the Scottish countryside charity APRS, has been researching the growth of hyperscale and AI datacentres across Scotland. She said that the claimed plans for 1GW of on-site renewable energy at the Lanarkshire site 'are total pie-in-the-sky. Where is this going to come from?' Jones added that the hyperscale datacentre would have the energy demand of half a million homes, an eighth of the entire Scottish electricity demand.
The government's plan to deliver the infrastructure the UK needs and get it connected has been met with skepticism. Rikap said that the rules are 'very flexible' and help companies to make big claims and investments that governments can use for their favor. The UK example will raise fears that, without oversight, the value of these investments may simply enrich companies and investors that are largely headquartered in the US.
Labor Market Data and Economic Implications
The UK government's AI plans center on creating thousands of jobs and constructing new datacentres. However, the true number of jobs created remains unclear. CoreWeave said that any job projections that have been shared in relation to these UK datacentre efforts have originated from the UK Government and DataVita, not CoreWeave. The government said that datacentres are vital to delivering the benefits of AI, from boosting productivity to improving public services.
The UK's AI drive is part of a global trend, with countries worldwide seeking to conjure economic growth from AI's transformative potential. With over £500bn promised in 2025, the value of these investments may be overstated. The UK government's plan to 'mainline AI into the veins of the economy' may be built on shaky ground, and the true cost of these investments remains to be seen.
