Skip to content
You can now search across every topic, entity and event.What's new
Data Centres: Boom and Backlash
14AUG

OpenAI puts a number on UK electricity gap

4 min read
10:48UTC

OpenAI cited UK industrial electricity at 'more than four times' US, Finnish, Norwegian, and Swedish rates as a primary driver of the Cobalt Park pause, the first time a hyperscaler has named energy cost as a co-equal barrier alongside the grid queue.

IndustryDeveloping
Key takeaway

OpenAI's 4x figure makes UK energy cost a permanent siting blocker until the merit-order pricing structure is reformed.

OpenAI publicly attached the multiplier 'more than four times' to UK industrial electricity rates relative to US, Finnish, Norwegian, and Swedish equivalents on 27 April 2026, naming the differential as a primary driver of the Cobalt Park pause in North Tyneside . The figure aligns with IEA price data showing UK industrial rates near $0.20-0.22/kWh against US $0.06-0.07/kWh and Finnish $0.07-0.09/kWh. For a 100 MW campus running at typical AI-training utilisation, that gap means roughly $100 million in additional annual electricity cost compared with a US site, before any consideration of compute density or cooling overhead.

OpenAI quantified what hyperscalers had previously said off the record. Until this week, the UK's headline data-centre problem ran through the grid queue: 50 GW of contracted demand against 45 GW of national peak . Hyperscalers had been treating energy cost as a secondary issue subordinate to interconnection wait times, on the working assumption that NESO and the AI Growth Zone designations would ease the queue. OpenAI's intervention separates the two and quantifies the second. With a queue solution, the UK still loses the campus economics by a factor of four.

UK industrial electricity pricing is shaped 60 to 70 per cent by gas-marginal merit-order pricing, plus carbon and network charges. None of those components have been reformed since the AI boom began, and the DESNZ and NESO policy responses available within the next twelve months do not change the merit-order structure. The implication is that the UK's data-centre future is becoming a talent-and-research base rather than a compute base; the capital that would have built Cobalt Park, Nscale's Glasgow expansion, and the Nottinghamshire AI Growth Zone is now redirecting into Nordic, Aragón, and Abu Dhabi tracks. The 50 GW grid queue persists, but it has been joined by a structural cost ceiling no rule cycle can lift.

Deep Analysis

In plain English

OpenAI explained publicly that electricity for businesses in the UK costs more than four times what it costs in the US, Finland, Norway, and Sweden. For a large data centre using 100 megawatts of power, roughly what a mid-sized city hospital district uses, that difference adds up to about $100 million in extra costs every year compared to a US location. This is the first time a major tech company has put a specific number on the energy cost gap and tied it directly to a decision not to build in the UK. The gap exists because UK electricity bills for businesses include several layers of charges for subsidising renewable energy and maintaining the grid that do not appear in Nordic or US tariffs at the same scale.

Deep Analysis
Root Causes

UK industrial electricity rates have three distinct structural components that do not appear in Nordic or US rates at comparable magnitude.

Network charges account for roughly £60-70/MWh of UK industrial tariffs in 2026 because the UK's Contracts for Difference (CfD) subsidy mechanism for renewables is recovered through network tariffs rather than direct taxation. The CfD mechanism has underwritten approximately £90 billion of offshore wind investment since 2014 and its cost recovery is embedded in the tariff structure until the early 2030s.

Balancing costs (the cost of keeping the grid stable when intermittent generation and demand are mismatched) add another £15-25/MWh in a market with 35 per cent wind and solar penetration but limited grid-scale storage. The UK's balancing mechanism is more expensive per MWh than any comparable European market because the GB grid runs at higher residual demand variability than the larger continental ENTSO-E synchronous zone.

The third component is the absence of a co-payment mechanism for transmission reinforcement: in the US, large loads trigger a generator-style interconnection study that assigns upgrade costs to the triggering party, capping unexpected network cost socialisation. In the UK, transmission upgrade costs are socialised across all users before cost recovery flows into network tariffs, which means existing users partly subsidise new entrants' connection costs, inflating the aggregate tariff.

What could happen next?
  • Consequence

    The UK falls off the global top-10 shortlist for new hyperscale greenfield sites as long as the 4x cost gap persists, redirecting an estimated 2-4 GW of announced UK data centre pipeline to Finland, Spain, and the Gulf before 2028.

  • Opportunity

    Ofgem's 2026 transmission tariff review and NESO's direct connection agreement process could, if they produce a 30-40 per cent reduction in network charges for large greenfield users, narrow the gap to 2-2.5x and restore UK competitiveness for mid-scale operators not requiring the Nordic price point.

First Reported In

Update #2 · Maine veto, Seattle freeze, $725bn capex

The Next Web· 6 May 2026
Read original
Different Perspectives
Antel, Uruguay's state telecom
Antel, Uruguay's state telecom
Antel will build a $70m AI data centre in Canelones, announced after an 11-12 August Council of Ministers meeting, with construction from Q1 2027 and roughly 250 direct jobs. Uruguay is putting a sovereign AI facility on the state balance sheet rather than courting a hyperscaler to build one.
Comtel
Comtel
Comtel is building a Figline Valdarno data centre inside the H2-Era Green Valley complex, powered by an 84 MW solar plant and selling waste heat to on-site farming. Co-location with generation and a heat offtaker is the model regulators elsewhere are now writing into connection terms rather than negotiating case by case.
Greg Abbott, Texas Governor
Greg Abbott, Texas Governor
Abbott directed PUCT and ERCOT on 3 August to audit every queued data centre's ownership, incentives and water use, saying projects that fail the state's survey should lose grid access. QTS's co-chief executives welcomed the review; Agriculture Commissioner Sid Miller called it too vague and wants a special legislative session instead.
Government of Gujarat
Government of Gujarat
Gujarat notified a Data Centre Policy on 7 August offering capital, interest, tax and duty subsidies for projects sourcing at least 51% green power, targeting 7.5 GW against a 74 GW installed base. Where other jurisdictions priced or restricted grid access this month, Gujarat is paying developers to come instead.
Taiwan's Legislative Yuan Economics Committee
Taiwan's Legislative Yuan Economics Committee
The committee gave preliminary approval on 5 August to a self-generation mandate for large power users, including data centres, with fines up to NT$750,000. It agreed the obligation in principle while leaving the megawatt threshold that defines its reach to later implementing regulation.
Australian Energy Market Commission
Australian Energy Market Commission
AEMC advised ministers on 5 August that data centres should surrender renewable certificates against actual consumption, prove new demand with new firm capacity, and register as market participants. The advice would turn passive customers into obligated participants inside the market's dispatch and settlement machinery, ahead of an access-standards determination due in late October.