AI Data Centers Are Straining the U.S. Power Grid. Federal Regulators Are Stepping In.
The Federal Energy Regulatory Commission ordered grid operators in June 2026 to reform how they connect large power users like AI data centers. About 30 to 50 percent of planned U.S. data centers face delays due to shortages of transformers and grid connections. Hyperscalers are projected to spend over 700 billion dollars on infrastructure this year.
The Federal Energy Regulatory Commission ordered regional grid operators in June 2026 to reform their interconnection processes for large power users, including AI data centers. The order came as the energy demands of artificial intelligence infrastructure pushed the limits of the U.S. electrical grid.
Hyperscalers, including Amazon, Microsoft, Google, and Meta, are projected to spend more than 700 billion dollars on data center capital expenditures in 2026. That spending is driving intense demand for electricity, cooling systems, and transmission infrastructure.
About 30 to 50 percent of planned U.S. data centers for 2026 face delays or cancellations due to shortages of transformers, batteries, and grid connections, according to industry analysts. In some regions, utilities have told data center developers they cannot connect to the grid for years.
FERC's order requires grid operators to justify or revise how they prioritize connection requests. The commission also encouraged operators to give preference to projects that use clean energy sources.
SoftBank announced a 75 billion euro commitment to build AI data center infrastructure in France, partly to reduce European dependence on U.S. cloud providers. France separately launched a 13 billion euro initiative to support domestic AI and deeptech startups.
Senator Bernie Sanders proposed in June 2026 that the U.S. government take a 50 percent equity stake in AI companies with annual revenues above 200 million dollars, arguing that the public should share in the wealth generated by AI systems built on publicly funded research.
