AI Data Center Boom Strains Global Power Grids
Artificial intelligence is driving a massive expansion of data center infrastructure, causing global electricity demand to surge and sparking local opposition across the United States.
Global electricity demand for artificial intelligence and data centers is surging, with the International Energy Agency projecting demand could reach 945 terawatts by 2030. In the United States, data centers accounted for 4.4% of electricity consumption in 2024, but projections suggest they could consume up to 20% of total production by 2035. This growth is driven by the high power requirements of Large Language Models and GPUs, which has shifted industry priorities toward speed of power access and delayed roughly half of the projects slated for 2026.
To support these workloads, Meta Platforms Incorporated, Google, Oracle, and xAI are constructing massive campuses across states including Texas, Indiana, and Ohio. Notable developments include Meta's $50 billion Hyperion campus in Louisiana and the Stargate initiative—a collaboration between OpenAI, SoftBank, and Oracle—deploying high-capacity sites in Texas, Michigan, and New Mexico. Fermi America is pursuing Project Matador in Texas, which seeks up to 17 gigawatts of private power.
The expansion has triggered significant local opposition and legal challenges regarding noise, property values, and environmental impacts. While some regions like New York have paused large-scale construction and San Marcos, Texas, has banned new facilities, President Donald Trump has advocated for easing regulations to allow data centers to build their own off-meter power plants to mitigate grid strain and rising utility rates for residents.
