OpenAI pauses Stargate in the UK, citing the energy costs and regulatory environment; OpenAI announced Stargate UK in September 2025 alongside Nvidia and Nscale
OpenAI said it is pausing its Stargate artificial intelligence infrastructure project in the UK, as it reins in ambitious spending plans ahead …
BloombergMark Bergen
Context & Ripple Effects
The UK project had been contemplated as part of Stargate’s international expansion, while OpenAI went ahead with a first European Stargate deployment in Norway. The pause therefore separates the feasibility of individual national builds from the broader European ambition.
It also follows reports that Stargate’s rollout was strained and that OpenAI shifted toward renting more AI servers from cloud providers. Energy cost and regulatory exposure now appear to be practical constraints on the UK leg, not merely planning variables.
First-order effects
OpenAI halts progress on the UK Stargate project, leaving Nscale and Nvidia without the expected near-term UK infrastructure deployment described in the announcement.
OpenAI’s UK compute plans must remain dependent on alternative capacity arrangements while the project is paused.
Second-order effects
Cloud providers and other available server suppliers become relatively more important to OpenAI’s UK-facing compute strategy, consistent with its reported move toward rented capacity.
UK data-center developers and power stakeholders face a clearer signal that AI-project demand is contingent on energy economics and regulatory certainty, rather than on chip demand alone.
Third-order effects
If similar pauses recur, frontier-AI infrastructure will be allocated more heavily toward jurisdictions that can pair large power supplies with predictable permitting and regulatory terms.
The pattern would make AI capacity a utility-and-policy coordination problem: model companies can distribute workloads, but dedicated campuses require durable local social and regulatory support.
The trend: Stargate’s UK pause is part of the shift from headline AI infrastructure commitments to country-by-country execution constrained by power, permitting, and access to rentable compute.
UK politicians should not be prostrating themselves at the altar of the eminently untrustworthy AI sector. Now they just look stupid...on all sides of the House...
Nearly half of US data centers planned for this year are expected to be delayed or canceled. Not a capital availability or chip problem. It's transformers and electrical equipment. Lead times have gone from 24 months to five years. The big four are spending $650B+ in 2026 alone […
Really disappointing to see, but I cannot say I am surprised. There was so much promise with AI growth zones and talk of energy abundance - not to mention the bravery in those wanting to change our outdated copyright rules. Seems like on all those issues we have failed to gain
Good. We need to stop these data centres to halt the Tech barons AI rollout. — But it forces the question what the government got out of these trade deals? As with the Pharma deal, an open ended commitment to be bullied by big business? — www.bloomberg.com/news/article...
another big infrastructure project canned by OpenAI ahead of its reported IPO...I need to put together a list of all the cancelled investments! — www.bloomberg.com/news/article...
Ah, the foul aroma of the tech industry trying to blackmail a government. “We have to take our imaginary toys away because your regulations make us sad.” — It's almost as if the deal was BS from the start. — www.theguardian.com/technology/ 2...
I'm old enough to remember when a US investment in the UK AI industry was announced in a flurry of publicity last September, worth £31bn. — Except it turned out to be one of those AI hallucinations you now hear about. — #AI #OpenAI — www.theguardian.com/technology/ 2...
There are plenty of Europe based and Digital Services Act compliant alternatives - link follows. www.theguardian.com/technology/ 2... www.theguardian.com/technology/ 2...
The Open AI/ Stargate UK story is rather intriguing. It was always rather vague... no way that OpenAI did not know about high UK energy costs when it was originally announced (though this clearly is a material issue) Only two things have changed in recent months. A row back on
I still believe, believe more than ever tbh, that compute is a commodity. I am completely unconcerned about OpenAI's imaginary AI data centre ambitions in the UK evaporating into nothing. https://tomforth.co.uk/...