The artificial intelligence boom is reshaping America's electricity landscape, with data centers now seeking power loads that rival entire cities. This unprecedented demand surge is forcing a once-in-a-generation reckoning for utilities accustomed to predictable, incremental growth.
The debates are unfolding at the nation's largest grid operator, PJM, and the Federal Energy Regulatory Commission (FERC). At the core of the controversy is a fundamental question: who should pay for the massive new infrastructure needed to support AI-driven computing?
Some proposals would allow data centers to connect directly to power plants or generate their own electricity on site, effectively operating outside the broader grid. These arrangements, while potentially faster to deploy, raise concerns about cost shifting and grid reliability.
The decisions will influence electricity prices, grid reliability, and the pace of AI development nationwide. A key FERC ruling could come as soon as this month, though multiple critical decisions are expected to unfold over months and years.
Critics argue that allowing data centers to bypass the grid could leave residential and small business customers footing the bill for infrastructure they don't use, while straining grid stability.