AI Drives Global Surge in Data Center Electricity Demand

Artificial intelligence is driving a global surge in electricity demand for data centers, ending a long period of stagnant power consumption in the United States and creating new pressures on energy grids. According to the Energy Institute’s 2026 Statistical Review of World Energy, the United States now consumes approximately 39.7% of the world's data center electricity, with its consumption rising to 312.6 TWh in 2025. Global demand reached 787.8 TWh that year, a 92% increase since 2020, with China and Europe following the U.S. as the largest consumers.

This growth is causing significant regional strain, particularly in Northern Virginia, where transmission congestion costs on the PJM grid rose 43% to $6 billion in the first half of 2026. S&P Global Energy forecasts that worldwide demand could reach 1,550 TWh by 2030, suggesting that energy availability may become the primary constraint on AI deployment.

Similar trends are appearing in Australia, where the Australian Energy Market Operator projects that data center energy demand will increase nearly sevenfold over the next decade, rising from five TWh in 2025/2026 to 34 TWh by 2035/2036. While the reliability outlook has improved, state and federal leaders are now discussing regulations for data center energy and water use. This follows previous decisions by the Queensland and Northern Territory governments to reject mandates requiring large data centers to invest in their own renewable resources.