The International Energy Agency (IEA)—widely regarded as the world's most authoritative energy forecaster—has put hard numbers on one of the technology sector's most discussed trends: artificial intelligence is on track to become one of the world's largest electricity consumers. In its April 2025 "Energy and AI" report, the agency projects that global data center power consumption will more than double by 2030, rising from roughly 415 terawatt-hours (TWh) in 2024 to approximately 945 TWh.

For context, 945 TWh is close to the combined annual electricity consumption of France and Germany, and would account for nearly 3% of total global electricity demand—up from a considerably smaller share today.

The numbers behind the surge

Data center electricity use has grown at an annual rate of about 12% since 2019. The IEA expects that pace to accelerate to roughly 15% per year through the end of the decade, with the growth driven almost entirely by AI workloads running on accelerated servers rather than traditional computing tasks.

The geographic concentration of the increase is pronounced. The United States and China are together forecast to account for nearly 80% of the total growth in data center electricity demand. The US alone is expected to add approximately 240 TWh of data center load by 2030—a 130% increase over 2024 levels—representing nearly half of all projected growth in total US electricity demand over the same period. In other words, data centers are on track to become one of the largest single drivers of new electricity demand on the US grid.

Capital is flowing to match. Data center spending reached roughly $500 billion in 2024, nearly double the level recorded two years earlier in 2022.

On the supply side, the IEA expects renewable energy to meet about 50% of the incremental data center demand by 2030. Natural gas and coal combined are projected to cover more than 40% in the near term, with nuclear contributing depending on regional availability—a mix that means the fuel blend serving AI infrastructure will vary meaningfully from market to market.

Why the grid is the real bottleneck

The report underscores a structural mismatch: data centers can be built in 18 to 24 months, while new transmission lines and generation capacity often require five to ten years to permit and construct. This timing gap is creating real stress on existing grids, particularly in regions such as Northern Virginia, central Texas, and parts of the US Southeast, where data center clusters are already straining local power systems. In practice, that means access to power—not physical construction—is emerging as the constraint that determines how quickly new AI capacity can be brought online.

The agency's findings point to an urgent need for expedited permitting processes and more diversified energy generation sources.

Market implications and the crypto dimension

The findings also point to a potential tension. If data centers absorb an increasingly large share of new power capacity, other electricity-intensive industries could face crowding-out effects. Bitcoin mining operations, which compete for many of the same resources—including low-cost power, grid interconnection, and favorable regulatory environments—may find themselves operating in a tighter market.

Several publicly traded Bitcoin miners have already begun pivoting toward AI hosting and high-performance computing, precisely because selling power-intensive infrastructure to AI customers can offer more predictable economics than mining tied to volatile crypto prices. Companies such as Core Scientific and Hut 8 have made this transition a central pillar of their corporate strategies, leveraging existing power contracts and data center shells to serve AI workloads. With the IEA's projections pointing to intensifying competition for electricity through the end of the decade, how permitting timelines, generation buildouts, and cross-sector demand for the same power resources evolve in major data center hubs will be a key dynamic to track.

Source: CryptoBriefing