Data Centers and the Limits of Federal Infrastructure Ambition

That fact is becoming one of the central legal and political constraints on the data-center sector. Federal officials can push faster permitting. Utilities can welcome large new loads. Investors can underwrite long-term demand. But none of those forces decide whether a particular project should receive priority access to a constrained grid, a finite water source, or a community’s remaining tolerance for industrial-scale development.

The question for sponsors is no longer simply whether a permit can be obtained. It is whether the project can hold a defensible claim to local resources and political acceptance long enough to be financed, built, and energized.

Federal Policy Cannot Control Local Siting Decisions

The federal government may view data-center development as a national economic and strategic priority. (See, e.g., here and here.) But data centers are built somewhere specific. They sit on particular grids, draw from particular water sources, and affect the communities around them. They may require new roads, substations, transmission upgrades, backup generation sources, utility expansion, and changes in how surrounding land is used and valued.

Those impacts are not abstract. Local officials must answer to residents concerned about noise, water use, electric rates, diesel generators, visual impacts, land consumption, and whether the promised benefits justify the burdens. State regulators must consider reliability, cost allocation, environmental impacts, and whether ordinary ratepayers should subsidize infrastructure for very large new loads. Utilities and regional grid operators must evaluate whether power can be delivered reliably and on what timeline.

While federal support matters, it cannot decisively resolve many tough questions. It does not allocate scarce capacity, assign the cost of grid upgrades, create available water, override local plans or guarantee political acceptance. Those questions are increasingly becoming the disputes that determine whether projects can actually be built.

New York Shows How a Pause Can Become Policy

New York’s recent data-center pause illustrates this shift. Governor Kathy Hochul’s Executive Order No. 62 temporarily pauses certain new hyperscale data-center projects while the state studies their cumulative effects on electricity demand, water resources, grid reliability, ratepayers, disadvantaged communities, and environmental quality. The order cites nearly 12 gigawatts of data-center load requests in the NYISO interconnection queue as of May 2026, with more than eight gigawatts entering the queue in 2025 alone.

For developers, the immediate concern is delay. The deeper concern is what emerges from it. A moratorium tied to agency study can create the record, standards, and policy architecture that govern future development. New York’s prior moratorium experience with high-volume hydraulic fracturing illustrates the point: a temporary review evolved into years of study, public participation, agency findings, and ultimately a durable policy framework. (For more, see here.) The analogy is procedural, not technological.

That makes a pause more than just schedule risk. It is rule-change risk. The review is not simply asking whether one facility should be approved; it is asking whether a rapidly growing infrastructure class needs new rules for power, water, ratepayer protection, community benefits and environmental review. New York shows the first move in that sequence: a project-specific pause becomes a system-wide inquiry. Texas shows the next move: that inquiry becomes a condition on access to the grid.

Texas Shows How Study Risk Becomes Project Risk

Texas presents the same dynamic in a different regulatory posture. Rather than pausing development through a statewide planning review, Texas is scrutinizing projects as they seek grid access. (See here and here.) Governor Greg Abbott has directed the Public Utility Commission of Texas (PUCT) and the Electric Reliability Council of Texas (ERCOT) to conduct a comprehensive verification and audit of data-center projects advancing through ERCOT’s interconnection process. The directive requires disclosure of power demand, self-generation capacity, water use and cooling technology, public financial assistance, ownership structure, and mitigation of community impacts. In the governor’s accompanying letter, he cited approximately 474 gigawatts of requests to connect to the Texas grid, with approximately 90 percent from data centers.

The significance is not administrative. Texas’s study requirements appear to make grid access contingent on reliability, cost allocation, water availability, community impacts, and whether the project will internalize infrastructure costs or shift them to ratepayers, utilities, or surrounding communities. In Texas, queue position will be less important than project viability.

The process may not end with approval plus conditions. It may produce a finding that a project should be redesigned, deferred, denied access to needed infrastructure, or abandoned altogether. That possibility means that a sponsor may have land under control, incentives negotiated, preliminary permits in motion and customers waiting, only to learn that the power timeline, water strategy, or local political record no longer supports the deal.

Local Moratoriums Are Not Just Local Politics

Recent local moratoriums show how quickly that concern can move from state policy to municipal control. Rochester Hills, Michigan recently extended a moratorium on data-center proposals while officials update zoning standards to address noise, vibration, light pollution, water requirements, electricity demand, and compatibility with surrounding uses. Other Metro Detroit communities, including Sterling Heights, Gibraltar, Pontiac, and Taylor, have reportedly adopted similar pauses.

That reaction is too easily dismissed as garden-variety NIMBYism. A community need not oppose AI or economic development to conclude that its zoning code is not equipped to handle data centers, which require unusually large electric loads, substantial cooling infrastructure, backup generation, and transmission upgrades.

Nor can federal policy easily erase those concerns. A locality asked to host infrastructure that affects local ratepayers, neighbors, water systems and land-use plans will want a voice in the terms. National demand for more data centers does not answer whether one jurisdiction should absorb the costs, risks and impacts of one facility.

The key point for sponsors is not that every local objection is persuasive. It is that legally relevant opposition often begins as local discomfort. If a concern can support a moratorium, an ordinance change, a site-plan condition, or a litigation theory, it belongs in the project risk model.

The Real Permit May Be Access to Systems

In that setting, the functional permit may be access to systems: power, water, and transmission capacity. Grid operators like PJM Interconnection may not formally approve or deny a project, but they can effectively determine whether and when it moves forward by controlling interconnection to the regional electrical grid. If transmission upgrades are delayed, queues grow, or grid constraints change the economics, the practical result may be indistinguishable from a permit denial.

Water presents a similar challenge. A project can obtain all necessary approvals and still face questions about water availability, cooling requirements, drought conditions, wastewater management, or competing local demands. Those questions are often controlled by different decision-makers than the ones responsible for zoning and land use.

Supply-chain delays compound the problem. Long-lead transformers, switchgear, cooling equipment, and transmission upgrades lengthen the interval between site selection and energization. The longer that interval, the more time opponents, regulators, utilities, and local governments have to change the assumptions on which the investment case depends.

For lenders and investors, these constraints increasingly resemble execution risk rather than traditional permitting risk. Site-control payments continue while queue positions slip. Utility-upgrade costs migrate from system planning into project economics. Water assumptions become politically unacceptable. Incentive packages become bargaining chips. For many projects, the formal land-use permit may not be the binding constraint. Power, water, transmission capacity, and community acceptance may be.

Opposition Becomes Legal Leverage

Residents may object to noise, emergency generators, water use, electric-rate impacts, land consumption, visual effects, transmission corridors, or the limited number of permanent jobs associated with some facilities. These community concerns can surface in zoning hearings, site-plan review, special-use proceedings, environmental comments, administrative appeals, development-agreement negotiations, utility proceedings, ballot politics, and litigation.

The same toolkit long used to oppose factories, pipelines, and renewable-energy facilities can be turned on data centers. Community concerns are not separate from legal risk. In contested infrastructure projects, opposition often becomes the mechanism through which legal vulnerabilities are identified and pursued. A dissatisfied community may not need to defeat the project on the merits if it can create delays, force new conditions, trigger additional review, or support a moratorium.

For sponsors, the issue is not simply whether a permit can be obtained; it is whether the project can survive the fight over scarce local systems. Questions that investors and lenders will ask before committing capital to data-center projects include:

• First, what has to be true for the site of a data center for a project to work? Land use, power, water, transmission, incentives, tax treatment, community acceptance, and customer timing are linked assumptions. 

• Second, who controls each project-critical resource? It is both traditional permitting agencies but potentially also utilities, grid operators, water authorities, local boards, state regulators, emergency-service providers and others with practical leverage over infrastructure access.

• Third, who pays if the overall energy system has to expand? Cost allocation can reprice a project as much as a permit condition can delay it, especially where upgrades are necessary to serve a single large load.

• Fourth, can the rules change before energization? Moratoriums, zoning amendments, utility rule changes, water restrictions, interconnection delays, new cost-allocation policies, and community-benefit demands should be treated as investment risks, not background noise.

• Finally, how can opposition to the data-center project impede the approval and permitting process? Community opposition that can support a moratorium, an ordinance change, a litigation theory, or a site-plan condition should be treated as a legal risk, not merely a political one.

Conclusion

AI will create new and unparalleled demand for electricity, transmission, water, and necessary supplies. As community resistance grows, governments will increasingly be asked to decide who receives access to scarce systems, on what terms, and at whose cost.

The data-center boom is becoming a test of who receives priority access to scarce local resources. The projects that succeed will not necessarily be those with the largest customers, the strongest incentives, or the fastest permitting timelines. They will be the projects whose sponsors can demonstrate, early and often, why the project deserves access to constrained local systems.