Data Center Moratoriums in 2026: What the Policy Shift Means for Developers
Fourteen states are now weighing restrictions on data center construction. The wave is structural, not cyclical.
On June 4, 2026, the New York State Legislature passed a one-year moratorium on siting, constructing, and commencing operations for new data centers. The bill awaits Governor Hochul's signature. The original proposal was three years; it was trimmed to one before passage. That the debate reached the floor at all tells you where this is heading.
New York is not an isolated case. As of June 2026, 14 US states have considered or are actively considering some form of moratorium or restrictions on data center development. Several have already moved to implementation. The pattern is the same in each: power consumption, ratepayer protection, and community opposition are converging into a political problem that no amount of tax incentives resolves.
What Is Driving the Moratoriums
Three forces are colliding simultaneously.
Power costs are rising visibly. PJM Interconnection's independent market monitor reported wholesale power costs in Q1 2026 were 76% higher than Q1 2025, with data center load cited as a key driver. In markets where ratepayers see their utility bills climb at the same time new data centers are announced nearby, the political equation turns hostile fast.
The employment argument no longer holds. Communities accepted earlier rounds of data center development partly because of job creation promises. Modern hyperscale and AI data centers require minimal ongoing staffing relative to their capital footprint and power consumption. That arithmetic is now widely understood by local officials.
Projects are outrunning grid capacity. According to analysis published in June 2026, approximately 8 billion in US data center projects have been blocked and 6 billion delayed since mid-2024 due to community and regulatory opposition. ConstructConnect data shows US data center construction starts reached 9.5 billion year-to-date in 2026, versus 3.6 billion in the same period of 2025. The gap between capital deployment and grid readiness is a structural condition, not a short-term mismatch.
What the Restrictions Look Like
State-level restrictions are not uniform. Some target scale: Oklahoma and Minnesota are focused on projects exceeding 100 MW. Some target cost allocation: Oklahoma's Data Center Consumer Ratepayer Protection Act requires large-load AI projects to fund their own grid infrastructure rather than pass costs to residential customers. North Carolina advanced legislation requiring utilities to absorb risk if AI and data center demand projections do not materialize.
The New York moratorium is the most sweeping. One year of no new siting, construction, or operations commencement. If Hochul signs, New York effectively exits the data center development market for the duration of the pause, and the precedent it sets for other states is significant.
Beyond state-level action, local moratoria are accelerating. The Town of Perth, NY passed a one-year local moratorium in June 2025. Virginia issued revised backup generator permitting guidance that challenges prior assumptions about how frequently hyperscale generators are actually used. Denmark's grid operator Energinet paused new grid connection agreements for large electricity consumers entirely.
The Developer Underwriting Implications
For institutional developers, these restrictions are not just political noise. They are site selection constraints that belong in the first pass of diligence, not the confirmatory phase.
Regulatory risk now precedes power risk in some markets. A site with strong grid access but in a jurisdiction trending toward moratorium is not a clean opportunity. The better underwrite identifies both constraints at the same stage.
The policy environment rewards early movers in clean markets. States and counties that have not yet triggered political friction around data centers retain the advantage. Secondary markets in the US Southeast, parts of the Midwest, and select Texas markets benefit precisely because they have not accumulated the community opposition that characterizes Northern Virginia or metropolitan New York.
Build-your-own-power strategies offer partial insulation. Oklahoma and similar frameworks that require developers to fund their own infrastructure costs shift the burden but do not eliminate it. A developer who brings a credible behind-the-meter power solution to a community conversation is in a fundamentally different position than one asking for grid upgrades funded by ratepayers.
The timeline risk is real. A one-year moratorium in a market where you hold an option means your feasibility clock is running, your interconnection queue position is aging, and your capital is tied up. Developers caught inside a moratorium zone with signed PSAs face extension negotiations, repricing, or abandonment.
The Broader Pattern
The 176 data centers permitted in the US in 2025 were the most in a single year on record. If all of them came online, their combined power draw would exceed every US state's consumption except Texas. That scale of demand, concentrated in the same grid regions where power prices are already rising, was always going to produce a political response.
What changes now is that the response has moved from community opposition to legislative action. The Sightline Climate estimate that 30 to 50 percent of the 2026 pipeline may be delayed or not delivered is consistent with this trajectory.
Developers who treat regulatory risk as a first-order site selection variable, not a late-stage permitting problem, are building a durable advantage. Those who do not are repricing it after the fact.