Six legal theories that could challenge state data center moratoriums

New York’s data-center fight is still in the awkward pre-litigation stage where the pleadings almost write themselves, but the winning theory does not. Executive Order No. 62 is already in place. S10642/A11560 has passed both houses, but as of July 25, 2026, it has not been signed by Gov. Kathy Hochul. The final version may still turn on thresholds such as 20 MW versus 50 MW, which matters because a moratorium that captures nearly every serious AI campus is a different litigation target from one that reaches only the largest new loads.[1]

That is why the most useful way to read the emerging New York data-center moratorium dispute over AI infrastructure regulation is not as a prediction that New York will lose. It is a risk map: which theories are colorable enough to appear in a complaint, which ones are mostly leverage, and which one could actually move the line between state siting authority and federal communications policy.

For readers who need the underlying New York framework before the litigation analysis, the companion New York data center moratorium legal analysis is the better starting point. This piece assumes the moratorium problem is real and asks what happens when counsel turns that problem into claims.

The queue growth explains the politics, not the legal answer

The legislative reaction did not come out of nowhere. NYISO’s large-load interconnection queue reportedly grew from 1,045 MW across 6 projects in 2022 to roughly 12 GW across 48 proposals by December 2025.[1] That is not a zoning-board inconvenience. It is a grid-planning event, especially when AI campuses arrive as concentrated, power-hungry loads rather than ordinary incremental commercial demand.

Nor is New York acting in isolation. MultiState’s 2026 guide tracked more than 200 data-center bills across all 50 states in 2025 and at least 14 states considering moratorium-type restrictions in 2026.[2] Foley & Lardner’s July 2026 analysis described New York’s move as a potential “turning point” and warned of a “first domino” effect for other states watching how far they can go.[3]

Those facts matter, but they do not decide the law. A state can have a legitimate energy-planning concern and still draft a restriction that collides with federal statutes. A developer can have a real business injury and still lose under rational-basis review. The work is in the gap between those two sentences.

Six theories are available; they are not equally dangerous

WilmerHale’s February 2026 client alert is the cleanest practice-ready map of the challenge theories now likely to be tested against state data-center restrictions. It identifies six possible routes: Equal Protection, Dormant Commerce Clause, Takings, Contract Clause, federal preemption under the Telecommunications Act, and additional federal statutory or administrative-law challenges depending on the final program design.[4]

That list is useful precisely because it separates complaint drafting from merits probability. Several of these claims are conventional enough to plead. Fewer are strong enough to reshape the case. One — Telecommunications Act §253 preemption — is genuinely disruptive if the FCC’s pending proceeding and New York’s final statutory language line up.

Equal Protection is usually a weak lead claim

Equal Protection will be tempting if New York’s final framework treats some large-load projects better than others. A 20 MW trigger, a 50 MW trigger, exemptions for certain industries, grandfathering for some applications, or different treatment for cryptocurrency, AI, cloud, or enterprise facilities could all invite a classification argument.

The problem is the standard of review. Data-center developers are not a suspect class, and large-load interconnection is not a fundamental right. Unless the final statute creates an unusually vulnerable classification, a court is likely to ask only whether the state had a rational basis for drawing the line. Grid reliability, emissions planning, ratepayer exposure, water use, and land-use sequencing are all plausible state interests.

That does not make Equal Protection useless. It can expose arbitrary exemptions and create discovery pressure if the legislative record is careless. But as a stand-alone path to invalidate a moratorium, it is likely to meet ordinary doctrinal resistance rather than judicial surprise.

Dormant Commerce Clause has more room, but it needs a record

The Dormant Commerce Clause claim is stronger than Equal Protection in one respect: it can focus on the practical operation of the moratorium, not only on its labels. A state rule that facially favors in-state economic actors, penalizes out-of-state infrastructure networks, or forces interstate digital services into an in-state planning bottleneck gives plaintiffs more to work with.

But a neutral moratorium keyed to megawatts, interconnection status, environmental review, or grid-impact study is not automatically discriminatory. The state will describe the law as a temporary tool for managing physical loads on a physical grid. Developers will answer that data centers support interstate commerce and that the burden falls on national cloud, AI, and telecommunications networks rather than on a purely local land use.

The case then becomes evidentiary. Did the state choose a threshold because of load impacts, or because it wanted to slow a disfavored industry? Are similarly situated high-load facilities treated differently? Does the moratorium function as a pause for study, or as an open-ended barrier to entry? How much interstate service capacity is actually delayed? The Dormant Commerce Clause can matter, especially when paired with a record of discriminatory purpose or excessive burden. It should not be inflated into a guaranteed winner.

Takings and Contract Clause claims are for hard retroactivity facts

Takings theories are most plausible where the moratorium reaches backward into projects that already hold land, permits, interconnection positions, or other concrete entitlements. A developer that merely hoped to build in New York has a different claim from one that bought land, negotiated utility arrangements, and proceeded under an existing approval path before the state changed the rules.

Even then, the doctrine does not treat every development delay as a compensable taking. Temporary moratoriums, study periods, and permit sequencing rules often survive when they leave economically viable use and serve conventional planning purposes. The stronger version of the claim would involve extraordinary retroactive application: a rule that strands a project after substantial reliance while offering no meaningful path to proceed.

The Contract Clause occupies similar territory. It is not a general-purpose weapon against regulatory change. It becomes relevant if the state substantially impairs existing contracts — for example, power arrangements, development agreements, or interconnection commitments — and if the impairment is not reasonably justified by the public purpose the state asserts. That is a project-specific inquiry, not a broad indictment of moratoriums as a category.

The §253 theory is the one that could change the boundary

Telecommunications Act §253 is where the analysis stops being a familiar land-use challenge and becomes something more consequential. Section 253 bars state and local statutes or regulations that prohibit or have the effect of prohibiting the ability of any entity to provide interstate or intrastate telecommunications service. WilmerHale flags the pending FCC proceeding on whether data-center siting restrictions can “materially inhibit” telecommunications services as the key federal backdrop for this theory.[4]

The move is not to say that every data center is a telecommunications carrier. The sharper argument is functional: modern data centers can be necessary infrastructure for communications services, cloud routing, network interconnection, and AI-enabled services that ride on telecommunications networks. If a state moratorium blocks the facilities required to provide those services at scale, the plaintiff may argue that the state has crossed from siting regulation into a federally preempted barrier.

That framing matters because it avoids the weakest version of the developer complaint. A court is unlikely to be moved by a generalized plea that AI competitiveness requires faster permitting. It may take more seriously a statutory claim that a state rule materially inhibits a federally protected communications function. The difference is not rhetorical; it determines whether the case is about economic inconvenience or federal preemption.

There is, however, a serious missing piece. The available authorities identify no circuit precedent directly applying §253 to data-center siting restrictions. That absence cuts both ways. It gives plaintiffs room to argue from first principles and from the FCC’s developing view of material inhibition. It also gives the state room to insist that traditional land-use, energy, and environmental regulation remain outside §253 unless the plaintiff can tie the moratorium to telecommunications service with much greater specificity.

The pending FCC rulemaking is therefore not background noise. If the FCC adopts a reading that treats certain data-center siting restrictions as materially inhibiting telecommunications services, plaintiffs will have a federal agency interpretation to place at the center of the complaint. If the FCC stays narrow, delays action, or draws distinctions that exclude ordinary AI compute campuses, the §253 claim becomes harder to plead as more than an extension of existing law.

The final New York text matters just as much. A statute that pauses only the largest new electrical loads pending defined agency review presents one case. A statute that effectively forecloses new facilities needed for covered communications services presents another. Thresholds, exemptions, grandfathering, duration, agency discretion, and the evidentiary findings accompanying the law could all affect whether the restriction looks like ordinary siting management or a material federal-service barrier.

Agency process may decide which claims survive the first motion

The remaining statutory and administrative theories are not free-floating objections to state regulation. They depend on the machinery New York actually builds. If the law delegates major decisions to an agency, counsel will look at notice, standards, timelines, evidentiary findings, waiver criteria, and whether similarly situated applications are handled consistently.

That is where seemingly technical drafting choices become litigation facts. A defined study period is easier to defend than an indefinite queue freeze. Published criteria are easier to defend than discretionary exceptions. A record explaining why a particular megawatt threshold captures relevant grid risks is easier to defend than a threshold that appears negotiated for political optics.

This is also where the legal-theory analysis differs from the broader zoning landscape. The existing piece on data center zoning legal challenges organizes disputes by plaintiff category — communities, developers, environmental challengers. A state moratorium case will often start with the developer or investor, but its legal center may be a federal-preemption theory rather than the ordinary local zoning record.

What counsel should watch before the first complaint

The litigation clock is likely to start quickly after enactment. The available analysis supports an expectation that the first lawsuits could be filed within 6–12 months of enactment, but the current answer remains contingent on the signed text, the implementing process, and the FCC’s §253 posture.[4]

• Final threshold: whether the signed law uses 20 MW, 50 MW, or another trigger, and whether that trigger sweeps in projects beyond the state’s stated grid concern.
• Grandfathering: whether projects with existing permits, land commitments, or interconnection positions receive a meaningful transition path.
• Agency discretion: whether the review process has standards, deadlines, and reviewable findings, or operates as an open-ended freeze.
• FCC action: whether the pending §253 proceeding treats data-center siting limits as capable of materially inhibiting telecommunications services.
• Legislative record: whether the state documents grid, emissions, ratepayer, and land-use reasons, or leaves evidence suggesting hostility to a disfavored interstate industry.

For AI infrastructure investors, the practical point is not that every delayed campus has a viable federal case. It is that the strongest challenge may not look like the usual zoning appeal. The claim with real disruptive potential is the one that characterizes the moratorium as a barrier to federally protected communications services, not merely as an unfavorable state energy policy.

Most familiar challenges face familiar limits. Equal Protection runs into rational basis. Dormant Commerce Clause requires a developed burden or discrimination record. Takings and Contract Clause theories need unusually strong retroactivity facts. Section 253 preemption is different: still novel, still untested against data-center siting, but capable of changing the legal boundary if the FCC and New York’s final statutory language give plaintiffs the materials to work with.

References

1. NY Executive Order No. 62 and S10642/A11560 legislative text, governor.ny.gov press release and Phillips Lytle analysis, governor.ny.gov press release and Phillips Lytle analysis
2. State Data Center Policy 101, MultiState, https://www.multistate.us/resources/state-data-center-policy-101
3. New York’s AI Data Center Moratorium: A Potential Turning Point for U.S. Data Center Development, Foley & Lardner, July 2026, https://www.foley.com/insights/publications/2026/07/new-yorks-ai-data-center-moratorium-a-potential-turning-point-for-u-s-data-center-development/
4. State Regulation of Data Centers: Emerging Trends and Potential Legal Complexities, WilmerHale, February 23, 2026, https://www.wilmerhale.com/en/insights/client-alerts/20260223-state-regulation-of-data-centers-emerging-trends-and-potential-legal-complexities