Senator Bernie Sanders of Vermont and Representative Alexandria Ocasio-Cortez of New York introduced legislation Thursday that would impose an immediate moratorium on the construction of new data centers in the United States, arguing that the staggering energy appetite of AI infrastructure was incompatible with the nation’s energy security needs and climate commitments at a time when the global energy system was already under extreme stress from the Iran conflict. The bill, called the Responsible AI Energy Act, was co-sponsored by eight other progressive members of Congress.

The legislation would halt the issuance of new permits for data center construction for a period of two years, during which the federal government would be required to conduct a comprehensive assessment of the energy and water demands of existing and proposed AI infrastructure, the capacity of the nation’s electrical grid to absorb continued AI-driven growth in demand, and the implications for the administration’s stated energy security priorities. New construction could resume after the moratorium period only under a new federal permitting framework that the bill would require Congress to establish.

Sanders and Ocasio-Cortez held a press conference Thursday at which they cited recent Department of Energy projections showing that data centers could account for as much as 12 percent of total U.S. electricity consumption by 2028, up from roughly 4 percent in 2023. The senator argued that allowing the AI industry to continue commandeering the nation’s electrical grid at the current pace, while ordinary Americans were paying record energy prices and the country was engaged in a war partly triggered by global energy competition, was a profound misallocation of national resources.

The legislation emerged directly from the convergence of two trends that had been building separately for months: the rapid expansion of AI data center construction by major technology companies including Microsoft, Google, Amazon, and a dozen large startups, and the global energy crisis triggered by Iran’s blockade of the Strait of Hormuz. Critics of the AI industry’s resource consumption had found a new and powerful argument in the energy emergency, and the moratorium bill represented their most ambitious legislative attempt yet to translate that argument into policy.

The AI industry reacted with alarm to the proposal on Thursday, with major technology companies and their trade associations issuing statements warning that the moratorium would devastate U.S. competitiveness in artificial intelligence and cede ground to China, which has been investing heavily in its own AI infrastructure. The Information Technology Industry Council said the bill was a reckless and shortsighted response to a temporary energy crisis that would have permanent and catastrophic consequences for American technological leadership.

The moratorium proposal also drew opposition from business groups, construction unions, and several bipartisan coalitions of state governors who argued that data center construction had become a major driver of economic development, tax revenue, and high-quality jobs in their states. Virginia, which hosts more data center capacity than any other state, sent a formal letter to congressional leaders Thursday opposing the moratorium. Several Midwestern governors argued that data center investment was transforming rural economies and should not be curtailed by federal action.

The political prospects for the bill were widely regarded as limited in the current Congress, where Republicans controlled both chambers and the administration had explicitly positioned the expansion of AI infrastructure as a national priority. The White House issued a statement Thursday criticizing the moratorium proposal as an attack on American innovation and economic growth that would benefit adversaries. Several senior administration officials had made data center expansion a centerpiece of their economic competitiveness agenda and had been actively facilitating private investment in the sector.

Nevertheless, the bill’s introduction was expected to have influence beyond its immediate legislative prospects. Congressional bill introductions by high-profile progressive figures like Sanders and Ocasio-Cortez frequently set the terms of policy debate and create pressure on both parties to articulate their positions on emerging issues. The legislation appeared timed to coincide with a growing public conversation about the resource demands of AI, a conversation that had been largely absent from the mainstream political debate just a year earlier but was now surfacing in kitchen-table discussions about electricity bills and energy prices.

The environmental dimensions of the debate added an additional layer of complexity. Data centers require not only enormous amounts of electricity but also vast quantities of water for cooling, a resource that is increasingly scarce in the American West and other drought-prone regions where technology companies have built many of their largest facilities. The Sanders-Ocasio-Cortez bill included provisions addressing water consumption alongside energy demand, and environmental advocates said the water issue deserved far more attention than it had received in previous congressional debates about AI infrastructure.

Energy economists offered mixed assessments of the moratorium’s potential effects Thursday. Some argued that a temporary pause would have little real-world impact because most of the data centers currently in planning and permitting stages would continue to be built once the moratorium expired, simply with a two-year delay. Others suggested that a moratorium could have a more lasting effect by forcing the industry to develop more energy-efficient technologies and operational practices, potentially changing the trajectory of AI’s long-term resource consumption. A third group of analysts argued that any government intervention in data center siting and construction would introduce costly regulatory uncertainty that would deter investment regardless of the specific rules ultimately adopted.

The debate over AI infrastructure energy demands was part of a broader reckoning that the technology industry was beginning to face about the physical constraints on its ambitions. The AI boom of the mid-2020s had been built on the assumption that computing power, energy, and capital would remain readily available and cheap, an assumption that the global energy crisis was now challenging directly. Whether the industry would adapt by developing more efficient systems, by accepting regulatory constraints, or by using its enormous political influence to push back against those constraints remained one of the most consequential open questions in American technology and energy policy.