Skip to content
You can now search across every topic, entity and event.What's new
Iran Conflict 2026
3JUN

OpenAI takes 88,500 sq ft in London, pauses Stargate UK

5 min read
09:04UTC

OpenAI signed an 88,500 square foot lease at King's Cross for more than 500 staff in the same week it paused its Stargate UK data-centre buildout, citing UK industrial electricity prices roughly four times US levels.

ConflictDeveloping
Key takeaway

Britain is becoming a research base for US labs while their training compute relocates to cheaper grids.

OpenAI signed a lease for 88,500 square feet at King's Cross in the week to 13 April 2026, framing the site as its biggest research base outside San Francisco and planning for more than 500 UK staff . In the same week the company paused its Stargate UK data-centre buildout. The pause was attributed in OpenAI's public statements to UK industrial electricity prices running at roughly four times US rates and close to twice French levels, with grid-connection delays of three to eight years against 18-24 months in comparable jurisdictions. The company said it would restart Stargate UK when regulation and energy costs permit long-term infrastructure investment.

The geographic split matters because The UK Sovereign AI Unit's policy design assumes research and training compute can grow together on British soil . OpenAI's decision fixes the opposite pattern: research in London, training compute in Texas. Frontier model training runs now consume GPU-hour volumes that make co-location of researchers and training clusters a commercial precondition, not a preference. A King's Cross office serves product and research iteration; it does not serve pre-training at frontier scale. The UK will recruit researchers; the training will happen where the grid can carry it.

A 3-to-8-year grid-connection queue is a binding constraint on UK sovereign compute that neither the Sovereign AI Unit's equity fund nor its separate £250m cloud procurement can compress. National Grid ESO connection timelines are set by network capacity and planning law, not by fund scale, and the Sovereign AI Unit is not a grid-investment instrument. That gap is the gap the fund has not yet addressed in public. DSIT's response has been to fund an AI Research Resource allocation per investee; OpenAI's response has been to lease an office and site the compute elsewhere.

A 500-staff London research office at King's Cross brings senior engineering roles, research output and a local tax base, and OpenAI's expanded UK headcount is not in itself a loss of sovereignty. The difficulty is that the Sovereign AI Unit was designed on the premise that British frontier-model capacity could be built on British compute, and a US lab's decision to pause its British compute buildout this week puts that premise under strain. The unit's infrastructure cohort will operate alongside a US lab that has already chosen the alternative answer to the grid-cost question.

Deep Analysis

In plain English

OpenAI, the company behind ChatGPT, has taken a large office in King's Cross, London, planning to hire over 500 people there. But in the same week it paused plans to build a major data centre in the UK. The reason is that electricity in the UK costs roughly four times what it costs in the US, and connecting a large data centre to the national grid takes three to eight years. Data centres training AI models consume extraordinary amounts of electricity, so those two problems together make the UK commercially unviable for the biggest AI infrastructure investments. The research office is coming; the power-hungry training computers are going elsewhere.

Deep Analysis
Root Causes

The UK's electricity cost and grid-connection disadvantage relative to France, Texas and the Nordic countries reflects three structural factors that compound each other.

UK wholesale electricity prices have been elevated since North Sea gas production peaked and declined from the early 2000s onward, requiring higher-cost gas imports to balance the grid. The Capacity Market mechanism, introduced after the Electricity Market Reform of 2013, adds a further per-MWh cost to industrial users to fund generation adequacy.

Grid-connection delays reflect a network designed for large centralised power stations connecting at a small number of points, not for distributed high-power consumers, data centres, battery storage facilities and heat pumps, connecting at hundreds of points simultaneously. Resolving any one of these three factors requires multi-decade investment programmes; resolving all three simultaneously while maintaining grid stability is the challenge National Grid ESO is working against.

What could happen next?
  • Meaning

    If OpenAI's model becomes the standard, research in the UK with training compute elsewhere, every British frontier-model ambition faces the same structural ceiling. DSIT's investees who build training infrastructure will either face uncompetitive operating costs on UK soil or will follow OpenAI's pattern and locate their compute outside the UK. At that point the sovereignty rationale for the Sovereign AI Unit weakens materially. The government's grid infrastructure commitments in the 2030 Clean Power Action Plan are the mechanism that could resolve this over a decade, but the Sovereign AI Unit's five-year investment horizon does not align with that timeline.

First Reported In

Update #3 · Sovereignty summit, minus the sovereigns

Tech Funding News· 23 Apr 2026
Read original
Different Perspectives
Oil markets and Lloyd's of London
Oil markets and Lloyd's of London
Brent fell to $89.25 on ceasefire probability, not new barrels, with traders voting for Trump's deed over Tehran's denial. Lloyd's has not repriced Hormuz war-risk cover because its trigger requires a UN Security Council resolution or government certification, so tanker insurance costs remain elevated regardless of the spot move.
Pakistan and Qatar mediators
Pakistan and Qatar mediators
Pakistan's Mohsin Naqvi was in Tehran for his second visit in under a week, using the Pakistan-Qatar channel that delivered April's ceasefire after an identical public-denial cycle. The channel carries both civilian and military buy-in from Islamabad, the only configuration Iran's split command cannot dismiss as a partial signal.
India
India
India summoned the US Deputy Chief of Mission after three Indian sailors were killed aboard MT Settebello, the first formal grievance from a major non-belligerent directed at US enforcement. Indian seafarers supply roughly 12 per cent of the global maritime workforce; their presence on third-flag Gulf tankers is structurally inevitable regardless of bilateral diplomacy.
Islamic Revolutionary Guard Corps (IRGC)
Islamic Revolutionary Guard Corps (IRGC)
The IRGC declared Hormuz closed on 11 June while civilian negotiators were on the same mediation channel, then issued no public comment on the MoU framework. Its silence on the framework, rather than any foreign ministry statement, is the operative approval signal; the corps' unilateral Hormuz closure shows it did not treat the diplomatic track as binding on its operations.
Iran foreign ministry (Baghaei)
Iran foreign ministry (Baghaei)
Esmail Baghaei told IRNA that reports of a finalised deal were 'merely speculation' and that Iran had 'not yet made a final decision'. The denial is structurally identical to Iranian foreign ministry statements during the April ceasefire talks, which produced a binding text within 48 hours of the same language.
Trump administration / CENTCOM
Trump administration / CENTCOM
Trump cancelled the third strike day and called the MoU 'very strong' and almost ready to sign, while CENTCOM kept tanker enforcement running in the same 24-hour window. The administration is simultaneously withdrawing the military pressure it claims drove the deal and sustaining the enforcement campaign it is trying to trade away.