The spread of “electricity-hungry data centers and clean-technology factories” throughout the United States is “leaving utilities and regulators grasping for credible plans to expand the nation’s creaking power grid,” according to a recent article in the Washington Post. According to the report, the surge in industrial power requirements due to such factors as innovation in artificial intelligence (AI) and the corresponding construction of large computing infrastructure, is prompting some officials to “look beyond the grid” to ease the “grid bottleneck” and meet electricity demands.

On December 20, 1951, researchers used energy produced by Experimental Breeder Reactor-I near Arco, Idaho, to illuminate four 200-watt lightbulbs. Since then, utilities have built commercial nuclear power plants in the United States almost exclusively to generate electricity. This has worked well alongside other power generation and transmission infrastructure—large oil- and coal-fired, natural gas turbine or hydroelectric plants, and a relatively simple electrical grid designed to deliver reliable power.

Humanity is now embarking on an epic and complex energy transformation across the grid, industry, and transportation. Renewables like wind and solar are contributing an increasing share of carbon-free electricity to the grid, but that contribution is variable and hard to predict—sometimes those sources produce more electricity than the grid needs, and sometimes less.