New York became the first state in the country to freeze new data center construction on July 14, when Governor Kathy Hochul signed Executive Order 62, a one-year moratorium on permits for facilities over 50 megawatts. Ohio’s Republican candidate for governor wants a pause too. So do Wisconsin’s and Florida’s Democratic candidates. Data centers are one of the only issues this year pulling both parties toward the same position — and the federal government is pushing the opposite way.
New York’s legislature passed the Responsible Data Center Development Act on June 4 — the Senate by 44-16, the Assembly by 102-39. Six weeks later, Hochul signed Executive Order 62, freezing discretionary state environmental permits for any data center of 50 megawatts or larger for one year. Projects already deemed complete before July 14 aren’t affected. Everything after it waits while the state writes new rules on utility rates, renewable energy use, and community benefits.
It’s the first statewide moratorium of its kind in the country. It isn’t the only one in motion — similar bills are pending or moving in more than 20 states.
U.S. data centers are on pace to nearly double their combined energy draw between 2025 and 2028, from roughly 80 gigawatts to 150, according to a January 2026 Bloom Energy report. Meta and Microsoft alone have committed roughly $725 billion to AI in 2026. That demand connects to the existing grid, and in places where data centers have clustered, the cost has shown up. In Virginia’s “Data Center Alley,” wholesale electricity prices in the most concentrated areas have risen as much as 267% over five years, per a Bloomberg analysis — a cost utilities pass to ratepayers over time. One Virginia resident’s monthly bill jumped from about $100 to $281 in a single month.
Water tells a similar story. A large data center can use up to 5 million gallons a day, comparable to a town of 10,000 to 50,000 people. About two-thirds of planned U.S. data centers are sited in areas that saw drought in the past year. Texas data centers alone drew more than 50 billion gallons of water in 2024.
Ohio Republican governor candidate Vivek Ramaswamy has called for a data center moratorium. So has Wisconsin Democratic candidate Francesca Hong. So has Florida Democratic candidate David Jolly, who said he’d support a construction pause “until the state has a responsible plan to protect our water, power grid, and communities.”
National polling shows the same pattern at different intensity. A Gallup survey conducted in March 2026 found 70% of Americans oppose a data center being built in their community, including nearly half who oppose it strongly. That opposition breaks 56% among Democrats, 48% among independents, and 39% among Republicans — genuine opposition on the right, just not as much of it. Asked why, 50% cited effects on local resources, 22% quality of life, 20% costs, and 16% pollution.
The opposition isn’t only rhetorical. More than $130 billion in proposed data center projects, at least 75 of them, were delayed or canceled in the first three months of 2026 alone.
On December 11, 2025, President Trump signed Executive Order 14365, “Ensuring a National Policy Framework for Artificial Intelligence.” It directs federal agencies to withhold broadband infrastructure funding and other discretionary grants from states with what it calls “onerous” AI regulations, and it creates a Department of Justice task force to challenge state AI laws the administration views as inconsistent with national policy. Reporting on the order’s text suggests it stops short of directly preempting state data-center-permitting laws themselves — but the funding threat is real, and it puts states weighing a moratorium in the position of choosing between local demands and federal money.
Supporters of continued buildout argue that AI leadership is now a function of physical infrastructure, not just software — that whoever builds and powers the most compute wins the broader competition, including with China. U.S. private investment in AI infrastructure, Meta and Microsoft’s combined $725 billion this year alone, already dwarfs the roughly $295 billion China’s government has committed over five years, and supporters say fragmented state and local permitting rules risk squandering that lead. The industry also points to the construction jobs, operating jobs, and local tax revenue large data center projects typically bring to host communities.
Opponents argue the costs of the AI boom are landing on people who see none of its benefits: electricity costs rising in data-center-heavy areas, water drawn from the same aquifers and rivers that supply farms and households, and permitting decisions moving faster than communities can weigh in. That a Trump-aligned Republican in Ohio and Democrats in Wisconsin and Florida are all calling for the same pause, they argue, shows this isn’t partisan positioning. It’s a shared complaint that new demand is being created and existing ratepayers are being asked to help finance the grid capacity it requires.
Neither side’s argument requires the other to be wrong. The country can want AI infrastructure and still not want its cost pushed onto residential electric bills. That’s the logic behind large-load tariffs, rate structures now approved in at least 24 states with four more pending, requiring data centers to pay the full cost of the grid capacity they use instead of spreading it across everyone else’s bill. It doesn’t resolve the water fight, and it doesn’t touch the federal funding threat hanging over states like New York. But it’s a working alternative to the choice between building everything and freezing everything, and more than half the country is already building it.
Hao’s reporting traces the human, environmental, and infrastructure costs behind the AI industry’s buildout — the on-the-ground version of the fight over megawatts and gallons now playing out in state legislatures.
Find it on Amazon →As an Amazon Associate I earn from qualifying purchases.