Gov. Kathy Hochul (D-NY) signed an executive order Tuesday imposing a statewide moratorium on the construction of large-scale AI and data centers.
The move marks the first time a U.S. state has paused such developments to evaluate how massive computing hubs affect the electrical grid and the environment. As artificial intelligence grows, the energy demands of these facilities threaten to strain existing infrastructure and increase costs for residents.
The pause applies to facilities with a maximum power usage of 50 megawatts or more [3]. This restriction will remain in place for one year [1], though some reports describe the duration as up to a year [2].
Hochul said the order is necessary to ensure that technology companies pay their fair share for electricity and to prevent further instability in the state's power supply. The governor said that the state cannot prioritize new industrial hubs over the needs of current citizens.
"I already have a shortage of power for the existing businesses and our hardworking New Yorkers who are paying through the nose in utility bills," Hochul said.
The administration intends to use the moratorium to conduct a comprehensive assessment of the environmental footprint of these centers. This includes analyzing water usage for cooling, and the overall carbon emissions associated with the high energy consumption of AI training models.
New York has seen a surge in proposals for data hubs, but the governor said that the state must first determine if the grid can sustain this growth without compromising reliability. The executive order effectively freezes new permits for the largest projects while the state develops a long-term regulatory framework.
“New York is the first state to impose a data-center moratorium.”
This executive order signals a shift in how state governments view the 'AI gold rush,' moving from unconditional recruitment of tech hubs to a cautious, resource-first approach. By linking data center permits to grid stability and utility costs, New York is creating a precedent that may lead other states to prioritize residential energy security over the rapid infrastructure expansion required by large-scale AI models.
