AI growth strains data centers and U.S. power grid

Schneider Electric presents an analysis of AI-driven data center energy challenges and proposed solutions.

• Main announcement/action: The report outlines that data centers currently consume ~4.4% of U.S. electricity, projected to rise to as much as 12% by 2028, with interconnection queues for new projects stretching to seven years and Dominion Energy projecting 5.5% annual demand growth in Northern Virginia (doubling by 2039) — requiring billions of dollars in infrastructure investment to accommodate AI-driven load growth.
• Background and technical details: The piece lists concrete mitigation options — DCIM, microgrids, liquid cooling, high-efficiency power distribution, small modular reactors/advanced fuel cells — and quantifies benefits such as digital substations improving grid capacity by 10%–30%, liquid cooling up to 3,000× more efficient than air cooling, and that demand response could reduce U.S. peak demand by up to 20%, while noting incompatible DR program rules, long permitting/interconnection timelines, and slow renewable deployment as key barriers.