IEA Forecasts AI Data Center Power Demand to Double by 2030
The IEA projects AI electricity use to hit 945 TWh in 2030, driving growth for fuel cell, nuclear, and utility firms.
Key points
- IEA projects AI data center electricity demand will rise from 415 TWh in 2024 to 945 TWh in 2030.
- Bloom Energy started 2026 with a $6 billion product backlog and a $14 billion service backlog.
- Cameco expects uranium supply to fall short of nuclear fuel demand by the mid-2030s.
The International Energy Agency projects that electricity consumption from artificial intelligence data centers will more than double between 2024 and 2030. This demand is expected to surge from 415 terawatt-hours to 945 TWh, a volume exceeding Japan's current total national power usage. The IEA, widely regarded as the world's top energy watchdog, highlights this trajectory as a structural shift in global energy requirements.
Data center electricity use has grown at an annual rate of 12% since 2017, significantly outpacing the broader rate of consumption growth. This rapid expansion creates a multi-layered opportunity across the power sector, benefiting companies that provide off-grid solutions, nuclear generation, and regulated utility services. The scale of this demand is reshaping investment strategies for energy providers.
Bloom Energy Secures Massive Backlog
Bloom Energy is positioned to capture demand that regulated utilities cannot immediately serve through its hydrogen fuel cell technology. The company entered 2026 with a product backlog of $6 billion, representing a 140% increase over its starting backlog in 2025. This figure excludes a separate $14 billion service backlog, which provides recurring annuity-like income through long-term maintenance contracts attached to each fuel cell sale.
Nuclear Operators Target AI Workloads
Constellation Energy, one of the largest U.S. nuclear operators, is actively partnering with technology giants like Meta and Microsoft to supply reliable power. Unlike regulated utilities, Constellation can charge market rates for its electricity, offering a competitive advantage in serving high-demand AI workloads. This strategy capitalizes on the sector's need for clean, baseload energy sources.
Uranium Supply Faces Tightening Constraints
Cameco, a major Canadian uranium producer, expects nuclear fuel demand to outstrip supply by the mid-2030s. This potential shortage could enhance pricing power for the company, which operates in a politically stable jurisdiction. The tightness in the supply chain underscores the critical role of fuel production in supporting the expanding nuclear fleet required for AI infrastructure.
NextEra Expands Data Center Reach
NextEra Energy is strengthening its position in the AI power market through its pending acquisition of Dominion Energy. This move grants NextEra access to Dominion's monopoly in one of the world's largest data center regions. The company maintains a dividend yield of over 3% and a history of annual dividend increases, appealing to investors seeking stability alongside growth in the AI energy sector, as noted by The Motley Fool.
