Global data center electricity consumption could surpass 1,000 terawatt-hours in 2026, according to International Energy Agency projections, with the United States accounting for more than 250 terawatt-hours of that total. The figures mark a steep climb from recent levels and reflect the rapid expansion of artificial intelligence workloads.

In 2024, data centers consumed roughly 415 terawatt-hours, about 1.5 percent of global electricity. Consumption has grown at about 12 percent per year over the prior five years, a pace that is accelerating as AI servers proliferate. The agency base case projects global data center demand will reach around 945 terawatt-hours by 2030, just under 3 percent of total global electricity.

The growth is concentrated in high-performance computing. Electricity use by accelerated servers, driven mainly by AI adoption, is projected to grow about 30 percent annually in the base case. Conventional server electricity consumption is rising more slowly at roughly 9 percent per year, underscoring that AI-specific hardware is the primary force behind the surge.

The data frames a structural challenge for power systems. As data center load grows faster than overall electricity demand, utilities and grid planners face pressure to add generation and transmission quickly. The concentration of new capacity in specific regions intensifies local strain, since a single large campus can require hundreds of megawatts of continuous power, reshaping demand forecasts well beyond historical patterns.

Source: International Energy Agency -- https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai