The United States has about 102.6 gigawatts of data-centre capacity operating or in the pipeline, according to Cushman & Wakefield figures published in a British Progress report on Britain’s AI strategy. That is 24.9 times the United Kingdom’s 4.12 GW, by our arithmetic.
The UK, Germany and France together come to 9.98 GW, 9.7% of the US total. The US pipeline alone, 70.63 GW under construction or planned, is 1.7 times the combined capacity of the next nine countries on the list, also by our arithmetic.
Short explainer · News Short
A vertical news Short narrated from this piece’s headline and opening line.
Business Pill · Sovereign AI
A one-minute explainer of sovereign AI: why countries want AI capacity under their own control. It teaches the general idea only and says nothing about any country’s figures in this piece.
The key insight: On Cushman & Wakefield’s figures, the US has about 102.6 GW of data-centre capacity operating or planned, 24.9 times the UK’s 4.12 GW, and its 70.63 GW pipeline alone is 1.7 times the combined totals of the next nine countries, by our arithmetic.
A chart of these figures, titled “Data-centre capacity by country, GW”, was posted on X on 29 October by @jujulemons with the comment “The UK and Europe’s current share of global AI compute is concerning enough. The trajectory is even more alarming.” The post had about 42,600 views when we read it on 30 October.
Its labels match the Cushman & Wakefield table that OfficeChai reported on 19 September, which describes the data as analysed by Computer Weekly and published in British Progress’s report “Dominant strategies for Britain in the age of AI”. Each country’s figure is capacity operational today plus pipeline additions, meaning capacity under construction or planned.
Data-centre capacity by country, GW, operational today and total including pipeline (under construction and planned): United States 31.97 to 102.60, China 5.49 to 8.15, India 1.58 to 5.11, Malaysia 1.29 to 4.88, Australia 1.49 to 4.82, Japan 1.94 to 4.46, United Kingdom 1.76 to 4.12, Canada 1.55 to 3.61, Germany 1.40 to 3.44, France 0.92 to 2.42. Source: Cushman & Wakefield data via British Progress, as reported by OfficeChai. The 1.7x ratio is our arithmetic.
Today Versus the Pipeline
The US has 31.97 GW operating today, per the table. Its 70.63 GW pipeline is 2.2 times that, so its total would be 3.2 times today’s capacity if every planned project were built, by our arithmetic.
China follows with 5.49 GW operating and 8.15 GW in total, then India at 5.11 GW, Malaysia at 4.88 GW, Australia at 4.82 GW and Japan at 4.46 GW. The UK is seventh, with 1.76 GW operating and 4.12 GW in total, ahead of Canada at 3.61 GW, Germany at 3.44 GW and France at 2.42 GW.
Where Europe Sits
The UK’s operating capacity is 5.5% of the US figure and its pipeline 3.3% of the US pipeline, by our arithmetic. The pipeline additions of the UK, Germany and France add up to 5.90 GW; the US pipeline is 12.0 times that.
Within Europe, the pipelines are proportionally similar: the UK’s total is 2.34 times its operating capacity, Germany’s 2.46 times and France’s 2.63 times, against 3.21 times for the US, by our arithmetic.
Against Britain’s Own Target
The government’s UK Compute Roadmap of July 2025 says: “We forecast that the UK will need at least 6GW of AI-capable data centre capacity by 2030 โ a threefold increase on the data centre capacity that is available in the UK today.” The roadmap was presented to Parliament by the science secretary.
Even if its whole pipeline were built, the UK’s 4.12 GW in the Cushman & Wakefield table would be 1.88 GW short of that 6 GW, by our arithmetic. The table counts all data centres, not only AI-capable ones, so the gap on the roadmap’s own definition may be larger.
What the Government and Builders Are Doing
On 28 October the Treasury published a Data Centre Covenant with Budget 2026, under which the government will use new powers to reserve and reallocate grid connection capacity for strategic projects, as we reported. The same Budget carried an AI package the Treasury put at £700 million.
The British neocloud Nscale features on both sides of the Atlantic. Microsoft pulled out of a UK data centre deal with Nscale at Loughton, Essex, Bloomberg and Reuters reported on 18 October, citing the Telegraph. Nvidia is in talks to back debt for Nscale’s 460MW Anthropic site in West Virginia, The Information reported, according to Reuters. For the chips going into these buildings, see our piece on Epoch AI’s compute data; for a European build, Scaleway’s AMD supercomputer.
The Structural Read
As we read the table, the gap is in the pipeline more than in today’s capacity: the UK operates 5.5% of the US figure but plans only 3.3% of the US pipeline, by our arithmetic.
By our arithmetic, the UK, Germany and France plan 5.90 GW of additions between them, and the US pipeline is 12.0 times that.
As we read the roadmap figure, even a fully built UK pipeline of 4.12 GW would be 1.88 GW below the 6 GW of AI-capable capacity the government forecast the UK will need by 2030, by our arithmetic.
UK Compute Roadmap, July 2025
“We forecast that the UK will need at least 6GW of AI-capable data centre capacity by 2030”
Three Implications
FOR UK POLICY The government’s own 6 GW forecast for 2030 is above the 4.12 GW the UK has operating or planned in the Cushman & Wakefield table, by our arithmetic.
FOR DATA-CENTRE BUILDERS The UK, Germany and France have 9.98 GW operating or planned between them, 9.7% of the US total, by our arithmetic.
FOR READERS OF THE VIRAL CHART Its labels match the Cushman & Wakefield table published by British Progress, which counts all data centres, not only AI-capable ones.
The analysis puts it this way: “The digital economy rests on a physical foundation that most people in that economy never see.”
As we read it, data-centre gigawatts are part of that foundation, and the Cushman & Wakefield table puts the US pipeline at 12.0 times what the UK, Germany and France plan together, by our arithmetic. The analysis does not discuss these figures.
What Is Not Established
Pipeline figures include planned projects, and some may never be built; the table does not say how much of each pipeline has power or permits. We read the table as OfficeChai reported it; we did not obtain the British Progress report or Cushman & Wakefield’s own release.
The quoted post by @SamuelAlbanie shows a different chart, a share of world compute based on chip arrivals, and we have not used its figures.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Go deeper (30 Oct): On Epoch AI’s customs-based estimates, the US held 72.2% of the world’s AI chip compute that had arrived since 2022 by mid-2026, up from 57.5% in the third quarter of 2023, by our calculation of the shares.
The Bottom Line
On Cushman & Wakefield’s figures, the US has about 102.6 GW of data-centre capacity operating or planned, 24.9 times the UK’s 4.12 GW, and the UK, Germany and France together hold 9.7% of the US total. Even fully built, the UK pipeline would leave it 1.88 GW below the 6 GW of AI-capable capacity its own roadmap says it needs by 2030, by our arithmetic.
A note on sourcing. Country figures are Cushman & Wakefield data as published in British Progress’s report and tabulated by OfficeChai on 19 September 2026; the 6 GW forecast is quoted from the UK Compute Roadmap (July 2025). Every ratio and sum is our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
