Data centre electricity on course to double by 2030: what the IEA reports

What we observe

In Energy and AI (April 2025), the IEA estimated data centre electricity consumption at around 415 terawatt-hours (TWh) in 2024, about 1.5% of global electricity consumption, having grown at 12% per year over the previous five years [1]. In its Base Case, the IEA projected that consumption would double to around 945 TWh by 2030, just under 3% of global electricity consumption, growing by around 15% per year from 2024 [1]. Electricity consumption in accelerated servers, mainly driven by AI adoption, was projected to grow by 30% annually in the Base Case, against 9% for conventional servers [1]. China and the United States account for nearly 80% of global growth to 2030; consumption rises by around 240 TWh (up 130%) in the United States and around 175 TWh (up 170%) in China [1].

For 2035, the IEA's sensitivity cases run from around 700 TWh (Headwinds Case) to above 1,700 TWh (Lift-Off Case), with the High Efficiency Case at around 970 TWh [1].

In Key Questions on Energy and AI (16 April 2026), the IEA reports that global data centre electricity demand grew by 17% in 2025, and that consumption from AI-focused data centres grew by 50% [2, 3]. The IEA's updated projections see consumption roughly doubling from 485 TWh in 2025 to 950 TWh in 2030, around 3% of global electricity demand, with AI-focused data centres tripling over the period [2]. The IEA says the central projection remains close to its 2025 trajectory, while bottlenecks across the value chain reduce the likelihood of more aggressive near-term scenarios [2].

Why it matters

Data centres remain a small share of global electricity, around 3% by 2030 in both IEA publications [1, 2]. Growth is concentrated in a few places: the IEA notes that data centres, unlike electric vehicles, cluster in specific locations, which can make grid integration more challenging [1]. The 2026 update adds a constraint: bottlenecks in grids, equipment and chips limit how fast capacity can expand in the near term [2, 3].

Method

Sources: two IEA web pages and an IEA press release, fetched and read on 26 Oct 2026 [1-3]. Figures and case names are as the IEA states them; the 2030 values come from different publications and are labelled by source. Chart data are in chart-data.csv. Framing follows AIO Watch method note AIO-M01, v1.2 [4].

References

1. [1] International Energy Agency. Energy and AI, "Energy demand from AI", April 2025. https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai. Accessed 26 Oct 2026.
2. [2] International Energy Agency. Key Questions on Energy and AI, "Executive summary", 16 Apr 2026. https://www.iea.org/reports/key-questions-on-energy-and-ai/executive-summary. Accessed 26 Oct 2026.
3. [3] International Energy Agency. Press release, 16 Apr 2026. https://www.iea.org/news/data-centre-electricity-use-surged-in-2025-even-with-tightening-bottlenecks-driving-a-scramble-for-solutions. Accessed 26 Oct 2026.
4. [4] AIO Research Team. AIO Watch method note AIO-M01, v1.2, 25 Oct 2026.