AI Demand Spurs US Gas Plants at Twice China's Pace

Governors Kathy Hochul of New York and Greg Abbott of Texas rarely see eye to eye, yet both have moved to slow the unchecked spread of data centers in their states. Hochul signed a year-long moratorium on new hyperscale facilities last month, citing the need to shield ratepayers from higher electricity costs. Abbott followed with a pause in Texas until regulators complete a full audit of existing projects and their impact on the power grid.

These actions reflect growing resistance across the country to the rapid construction of data centers built to support artificial intelligence. The facilities demand enormous amounts of electricity, much of it now coming from new natural gas plants. Analysis from Global Energy Monitor shows the United States is now developing twice as much gas-fired power capacity as China, with the total in development jumping 50 percent since January to 378 gigawatts. Roughly half of that expansion ties directly to data centers.

Tech companies have increasingly turned to private behind-the-meter gas plants to avoid delays connecting to the public grid. This approach has helped sidestep some public scrutiny, but it has not stopped local opposition. Communities worry about higher utility bills and the health effects of added pollution, including nitrogen oxides and other emissions from the plants. The Trump administration's Environmental Protection Agency has proposed changes that would make air pollution disclosures for these projects voluntary in many cases, removing the requirement for public notice under the Clean Air Act.

Bipartisan concern is mounting. Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez have backed a federal moratorium bill, while Senator Josh Hawley has supported measures aimed at preventing data centers from raising electricity rates. In Michigan, candidates are campaigning against the projects, and Pennsylvania Governor Josh Shapiro recently added strict new conditions after previously welcoming large developments.

The five largest hyperscalers plan to spend more than $750 billion this year on capital projects, with most of that directed at AI infrastructure. That scale has driven utilities to revise their forecasts for peak electricity demand sharply upward. Critics argue the benefits flow mainly to a handful of large corporations while the costs, in the form of dirtier air and strained power systems, fall on ordinary households.

Public pushback has already produced results at the state and local levels. Lawmakers in both parties now face pressure to impose tighter rules rather than continue the rapid approvals seen in recent years. The pattern suggests that without stronger safeguards, the expansion will continue to collide with communities unwilling to absorb the downsides.