Data centres’ electricity use doubles to 1,000 terawatt-hours by 2030 in the IEA’s Stated Policies Scenario, the agency said in its World Energy Outlook 2026 on 19 November.
The IEA wrote that “the boom is also revealing several constraints related to grids, supply chains and, in some cases, public acceptance”. Bloomberg summed it up as a data-centre boom pushing power grids to the limit.
Business Pill · Lead Time Explained
An evergreen Business Engineer explainer on lead times. It teaches the general idea only and does not discuss the IEA outlook.
The key insight: In its Stated Policies Scenario, the IEA’s World Energy Outlook 2026 sees data-centre electricity use doubling to 1,000 TWh by 2030 and names constraints related to grids, supply chains and, in some cases, public acceptance. In 2025 data-centre demand grew about 5.7 times as fast as overall demand, by our arithmetic.
What the IEA Projects
The outlook says “the boom in artificial intelligence-related investment is pushing up electricity consumption by data centres, which doubles to reach 1 000 terawatt-hours by 2030 in the STEPS”.
Data centres are one of several drivers. The executive summary lists electric motors, appliances, air conditioners, EVs, heat pumps “and data centres by the AI boom”.
Overall, global electricity demand grows by more than 10,000 TWh over the next decade, the IEA says. Electricity’s share of final energy consumption is “closing in on 30%” in two scenarios, up from 23% in 2025.

Update, 19 November 2026, 12:39 UTC: Bloomberg reports a $4 trillion data-centre investment outlook from the IEA; The Guardian reports the agency tracks 93 AI factories heading to about 50 GW. Global data-centre investment is expected to reach $4 trillion by 2030, Bloomberg reported, citing the International Energy Agency.
Bloomberg added that fewer than 100 major AI projects would need a combined 50 GW, matching peak UK demand.
The IEA used space-based imagery to track the buildout of 93 large “AI factories” worldwide, The Guardian reported. Its analysis suggests the world’s data-centre capacity has climbed tenfold in the last 24 months to 17 GW; once the AI factories are completed, their combined capacity is likely to be about 50 GW, the agency said.
The world’s existing data-centre stock stands at 120 GW, 80% of it in China, the US and the EU. A further 120 GW will be built by the end of the decade in the IEA scenario often treated as its central case, three-quarters of it AI data centres, The Guardian reported.
On power sources, there were plans in August for about 180 GW of new gas-fired plants linked to US data centres, against 11 GW in Canada and 4 GW from AI developers in the rest of the world, the IEA said. It sees data-centre emissions rising 70% by 2030.
By our arithmetic, 50 GW across 93 projects averages about 0.54 GW each, and the central case would double the stock to about 240 GW by 2030. The $4 trillion figure comes from Bloomberg’s summary only; The Guardian’s report does not give it.
How Fast Data Centres Are Growing
In April, the IEA said electricity demand from data centres “soared by 17% in 2025”, against 3% growth in global electricity demand (IEA). That is about 5.7 times as fast, by our arithmetic.
The same release said the capital expenditure of five large technology companies passed $400 billion in 2025 and was “set to increase by a further 75% in 2026”. On those figures, that is roughly $700 billion this year, by our arithmetic.
Where the Strain Shows
The IEA’s executive summary says “surging demand from data centres is amplifying strains in some markets and supply chains”, even as AI “could help to optimise the electricity system and bring down costs”.
On grids, the outlook says at least 25 million kilometres of new power lines are needed over the next ten years, and 15 million kilometres of existing lines need replacement or upgrades.
In April the agency also said the pipeline of data-centre projects was “holding up grid connections and other necessary approvals”, and that developers are advancing projects with onsite gas generation, “largely in the United States”.
The Local Fight
Public acceptance is already a live issue. Alberta has proposed a 600-metre buffer between data centres and homes (our report). We explain how local moratoriums work in How Local AI Bans Work.
Demand is being counted in gigawatts. Musk says Anthropic is eyeing 10 GW more from SpaceX’s AI cloud, as we reported in our Musk piece.
The Structural Read
As we read the outlook, the binding constraint on AI build-outs is shifting from chips to power lines and permits.
As we read it, the IEA’s own pairing of grid strain with public acceptance explains why local setbacks and bans now sit next to capex numbers.
IEA, World Energy Outlook 2026
“The boom is also revealing several constraints related to grids, supply chains and, in some cases, public acceptance.”
Three Implications
FOR DATA-CENTRE DEVELOPERS The IEA says the project pipeline is holding up grid connections and other approvals.
FOR UTILITIES AND GRID BUILDERS The IEA says at least 25 million kilometres of new lines are needed over ten years.
FOR BIG TECH CAPEX Five large tech companies spent more than $400 billion in 2025, set to rise a further 75% in 2026, per the IEA.
What Is Not Established
The pages we read give the 2030 figure only for the Stated Policies Scenario; we have not seen regional data-centre splits or a 2035 figure. We read only Bloomberg’s headline and post, not its article.
The Bottom Line
In its Stated Policies Scenario, the IEA sees data-centre electricity use doubling to 1,000 TWh by 2030, and it names constraints related to grids, supply chains and, in some cases, public acceptance. Data-centre demand grew about 5.7 times as fast as overall demand in 2025, by our arithmetic.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The 1,000 TWh figure, the grids-supply-chains-public-acceptance line, the 10,000 TWh, 30%, 23% and power-line figures and the executive-summary quotes are from the IEA’s World Energy Outlook 2026 pages published on 19 November. The 17%, 3%, $400 billion, 75%, grid-connection and onsite-gas lines are from the IEA’s April release on its Key Questions on Energy and AI report. We read only Bloomberg’s headline and X post. The 5.7 times and roughly $700 billion are our arithmetic.
Sources: IEA, ‘World Energy Outlook 2026: Overview and key findings’, 19 Nov 2026 · IEA, ‘World Energy Outlook 2026: Executive summary’, 19 Nov 2026 · IEA, ‘Data centre electricity use surged in 2025, even with tightening bottlenecks driving a scramble for solutions’, 16 Apr 2026 · Bloomberg, ‘IEA Sees Data Center Boom Pushing Power Grids to the Limit’, 19 Nov 2026 (headline and post only)









