Direct Answer: Treat a Data Center as Critical Infrastructure, Not an Ordinary Warehouse
Cities reviewing data center permits in 2026 should evaluate each project as a combined land-use, utility, environmental, and economic-development decision. A data center may resemble a large industrial building, but its electrical demand, water use, noise profile, freight activity, tax incentives, and effects on the regional power system can be unlike those of conventional warehouses. The strongest review process begins before the application is formally accepted, requires utility confirmation of available capacity, and prevents public subsidies or tax abatements from being promised until infrastructure impacts and community benefits are quantified. This approach matters because a single large campus can consume as much electricity as tens of thousands of homes, although actual comparisons depend on campus design, computing load, and the year measured.
Also worth reading: How Should Cities Use Responsible AI to Speed Up Building Permits Without Weakening Public Oversight? · Which Contract Clauses Should Cities Use for AI Permit Review Systems? · How Should Cities Evaluate AI Planning Tools for Safer, Faster Development Review?
There is no universal threshold at which a data center becomes “too large” for a city to permit. Instead, planners should examine several measurable triggers, including a proposed electrical load of 20 megawatts or more, expected water demand above a locally established threshold, transmission-line or substation work, acreage exceeding a designated employment or industrial district, or a request for tax abatements, impact fees, or infrastructure assistance. A jurisdiction may set lower triggers for environmentally sensitive areas or places with constrained grids. The governing principle is that exceptional infrastructure demands should produce extra review rather than automatic approval or rejection. Municipal standards should be objective, published in advance, and available to applicants and neighboring property owners.
How the Permit Review Should Work in Practice
A sound review starts with a pre-application meeting in which city staff identify permitted uses, environmental requirements, utility constraints, road access, and likely community concerns. The applicant should then submit a technical package containing site plans, building height and setbacks, equipment inventories, utility letters, construction schedules, and a financing description. For electrical service, planners should request a preliminary load study rather than relying on a utility service letter alone. That study should distinguish IT load from mechanical load and account for backup generators, chillers, battery systems, and future expansion phases. For water, the application should distinguish consumptive process water from water returned to the sewer or local watershed. These distinctions can materially change the project's apparent impact.
The planning department should coordinate its work with the electric utility, water provider, fire marshal, public works agency, transit or road officials, and relevant environmental regulators. Public hearings must address technical impacts without allowing technical complexity to obscure uncertainties. Staff should publish maps, modeling assumptions, mitigation commitments, and enforceable permit conditions. A development agreement may be appropriate when the applicant seeks zoning changes, road improvements, tax abatements, or public funding, but it should not replace compliance with ordinary planning, building, fire, environmental, and engineering requirements. Any community-benefit agreement should identify who receives money, who performs the work, when funds are paid, and how missed targets are corrected.
A complete review also needs a public record. Cities should preserve comments, staff reports, utility correspondence, studies, conditions, and approval votes. This is especially important where officials authorize a temporary pause or moratorium while new rules are developed. A pause should have a defined end date, written findings, and a path for pending and new applications. Otherwise, applicants may challenge the process as arbitrary, while residents may assume that the pause is a permanent prohibition. Clear records also help the city explain why two data center proposals with different grid, water, and fiscal effects received different outcomes.
Why Data Centers Require Different Permit Standards
Conventional industrial standards generally focus on building massing, setbacks, parking, loading, and compatibility with nearby properties. Those controls remain relevant to data centers, but they do not capture their infrastructure demands or operating conditions. Data centers may include generators, fuel tanks, cooling towers, electrical substations, battery storage, cooling-water systems, and transmission facilities. Their largest physical feature may therefore be off-site infrastructure. A parcel that meets every site-development standard could still strain the local grid or require a new substation, road connection, and drainage project. Accordingly, the permit file should review both the land and the infrastructure serving the land.
Data centers also differ from warehouses in operating hours, employment, traffic, and energy intensity. Many run continuously, generating mechanical noise, heat rejection, and nighttime light even when the building has few employees. Conversely, they can produce relatively little compared with manufacturing or freight-intensive industry, depending on efficiency and location. Policymakers should not classify the industry solely as either harmless or harmful. They should measure actual and projected demand, design quality, and mitigation. Efficiency labels, company claims, and estimates from promotional studies should be treated as inputs, not final evidence.
The public interest extends beyond project design. A city may decline a project if the responsible utility cannot secure reliable service without disproportionate public investment or if the project would worsen documented reliability risks. However, grid constraints are regional as well as local, and a city should not assume that one company's unsuccessful interconnection request proves the same outcome for every proposal. Transmission planning, data-center curtailment practices, on-site generation, battery duration, and phased construction can change feasibility. The city should ask the utility to explain queue position, expected service dates, upgrade responsibilities, and costs rather than reducing the issue to a binary statement that power is or is not available.
Comparison of Regulatory Options
| Feature | Use-by-right data center district | Conditional-use permit with infrastructure standards | Moratorium or project-specific review | Performance-based overlay with possible incentives |
|---|---|---|---|---|
| Approval speed | Usually fastest if standards are met | Moderate; adds evidence and hearing time | Slowest while rules or project review is pending | Variable and often negotiation-heavy |
| Certainty | High for compliant projects | Moderate if conditions are objective | Low during a temporary pause | Lower when incentives depend on later agreements |
| Community influence | Limited to code compliance | Public hearing and conditions | Strong during policy formation | Strong but potentially technical |
| Infrastructure response | Basic design and service review | Can require load, water, emergency, and fiscal studies | Can identify system-wide concerns | Can price impacts and link benefits to commitments |
| Fiscal risk | Lower if public costs are fully charged | Depends on conditions and legal authority | Can delay project revenue | Higher, but may be controlled through agreements |
| Best use | Predetermined suitable industrial areas | Areas with manageable uncertainty and defined standards | A short, time-limited need to study new impacts | Jurisdictions prepared to enforce measurable benefits |
| Main weakness | May not address cumulative effects | Can become predictable or developer-friendly if weak | Can increase uncertainty and litigation risk | Open-ended promises can be hard to enforce |
Practical Steps for Applicants, Neighbors, and City Staff
Applicants should prepare a “constraint-first” package rather than presenting finished architectural renderings before confirming basic feasibility. The first package should identify parcel boundaries, existing and proposed utility service, acreage, phasing, expected power and water demand, backup-fuel capacity, anticipated jobs, and any request for public assistance. It should also disclose ownership, financing, tenants or customers where known, and whether the facility may be converted from AI training to cloud computing. These changes can alter schedules and demand profiles. Neighborhood groups should request the same information independently and compare it with adopted plans, utility documents, and the application of comparable projects.
City staff should establish a published review matrix with named responsible agencies and target dates. For example, planning can review land use, public works can assess roads and drainage, the utility can address electrical service, and the fire marshal can review fuel, emergency access, and fire-water strategy. A planning commission can recommend conditions to the city council, but technical reviewers should first identify unresolved factual gaps. Officials should not force unsupported minimums into the planning process. If the applicant requests relief from a standard, staff should state the requested modification, its expected effect, any alternative compliance path, and who bears the associated public cost.
Neighbors should focus comments on measurable effects and enforceable remedies. Useful subjects include transformer hum, generator testing hours, cooling-tower drift, truck routes, visual screening, water consumption, emergency notification, and construction practices. Requests for a wall or buffer may not address low-frequency transformer noise, while a general promise to “minimize noise” may be unenforceable. A better condition may specify equipment limits, operating periods, monitoring locations, complaint procedures, and corrective deadlines. Residents should also recognize that data centers can contribute tax revenue, construction employment, and a comparatively low-waste industrial use, but those benefits should be compared with public costs rather than counted automatically in the project's favor.
Costs, Incentives, and Financial Risk
The city should distinguish private project costs from public costs. Applicant-funded improvements may include service upgrades, dedicated roads, water facilities, drainage, screening, and generator systems. Public costs can arise through tax abatements, impact-fee reductions, grants, bond issuance, utility make-ready work, or public maintenance of infrastructure serving the project. In a financially constrained municipality, a tax abatement can be material even when the project is privately financed. Applicants should provide a full fiscal model showing assessed value, exemption duration, expected abatements, jobs, wages where supportable, construction taxes, and public improvement costs. Gross property-tax potential should not be presented as net revenue without deducting exemptions, service obligations, and administrative costs.
A project-specific data center impact fee should reflect locally identified costs, not broad political perceptions. One potential structure separates a land-use review and building-permit fee from utility-capacity and emergency-response costs. Another divides obligations among the applicant, rate-paying customers, and taxpayers according to documented responsibility. Cities should not charge an arbitrary fee merely because data centers have become controversial. Conversely, recovering costs through general revenues may be unfair if the company receives special utility treatment or public incentives. Developers, utilities, and the city should document assumptions and update them when project phases change.
Incentives should be conditional and time-limited. For example, any reduction could require payment of a larger fee, on-site backup generation that does not rely on a limited emergency resource, or participation in a demand-response program. A contract might also specify local hiring, vendor procurement, water-efficiency targets, or restoration of a monetary community fund. These terms need baseline definitions, measurement methods, reporting dates, and penalties. A broad community-benefit agreement is weaker if the city cannot verify compliance. Fiscal impacts should also be evaluated under downside cases, including delayed construction, lower occupancy, reduced tax valuation, or withdrawal of the tenant. No credible cost estimate is precise enough to ignore these possibilities.
Common Mistakes That Produce Weak Decisions
A frequent mistake is treating “available power” as equivalent to “deliverable power.” Utility territory alone does not resolve substation, transmission, generation, or interconnection constraints. Applicants should provide a dated study showing requested capacity, available capacity, required upgrades, queue position, expected energization, and the effect of later phases. Some jurisdictions have responded to data-center growth through temporary pauses while officials study grid and infrastructure consequences, but a pause does not itself create power or set permanent standards. The eventual decision should rely on evidence and a lawful process.
Another error is relying on a technology-neutral use label that permits nearly any scale. A use called “digital infrastructure” may include a small edge-computing facility and a multi-building hyperscale campus. Standards should address scale through measurable features or thresholds, while allowing ordinary minor projects to use a less burdensome process. Cities should avoid invented precision, such as assuming that every data center will use a fixed amount of water per square foot. They should require project-specific estimates and require updates when cooling systems or workloads change.
Officials also make mistakes by promising jobs that the operator cannot control, treating tax exemptions as a permanent net gain, or allowing a development agreement to override previously adopted environmental standards. Public statements should separate construction jobs from permanent positions and distinguish direct employment from contractor or vendor activity. Communities should also be wary of the claim that data centers are always environmentally benign. Low direct water use can coexist with substantial electricity demand, and efficient new equipment can still add to regional load. Balanced analysis considers both project efficiency and cumulative growth.
When a City Should Act, Pause, or Decline
A city should revise its rules before the next significant application when its ordinance lacks provisions for substations, fuel systems, cooling equipment, emergency planning, or off-site infrastructure. It should act immediately when there is an unpermitted structure, unsafe access, fuel or fire-code issue, material noise nuisance, or unapproved clearing. A complete project review can proceed while the city develops broader performance standards, provided the existing ordinance authorizes it. A moratorium should be reserved for a documented gap that cannot be addressed through conditions on pending cases, and it should include an end date, public findings, and treatment of vested and pending applications.
Denial is appropriate where an applicant cannot satisfy objective legal requirements, such as access, fire safety, water availability, or a valid use permit. Officials should also consider denial or redesign when the responsible utility finds that service cannot be provided by an acceptable date, when public costs cannot be lawfully or fairly assigned, or when mitigation cannot reduce documented harm to a protected area. The city should not deny a project simply because officials dislike its industry or because residents demand a ban without evidence. Equally, a local preference for data centers should not override state utility law, constitutional constraints, environmental permits, or regional grid obligations.
The best decision is frequently conditional approval with firm phasing and infrastructure terms. That can allow useful development while reserving later phases until grid, road, water, and fiscal conditions are met. Conditions should include expiration dates and a rule that unused approval does not guarantee later construction. This is particularly important for projects whose first phase is modest but future campus demand is much larger. A phased permit can prevent the city from evaluating only the opening building while approving consequences for infrastructure that may not arrive for several years. As of October 2, 2026, cities should expect continued policy experimentation, including test applications, new zoning districts, project-specific conditions, and temporary pauses.
The Recommended Permit Decision Standard
The definitive standard is not whether a city likes or dislikes data centers. It is whether the jurisdiction can permit the project lawfully, reliably, and fairly while making its infrastructure costs and community effects visible. Cities should adopt clear definitions, separate minor facilities from major campuses, require technical studies at reasonable thresholds, and publish the evidence supporting each decision. They should give applicants a predictable path, neighbors a meaningful review role, and utility and emergency agencies a direct role in factual findings.
Ultimately, data center policy should connect every exceptional demand to a corresponding responsibility. Large electrical loads require credible service plans; water use requires metering and disclosure; generators require emergency planning; public incentives require fiscal accounting; and regional scale requires a process for monitoring cumulative effects. This method does not assume that data centers are universally beneficial or universally harmful. It recognizes that a data center can be compatible with a particular location while remaining unsuitable for another, and that the same operator or technology may produce different impacts across sites. For applicants and residents alike, transparent models, enforceable conditions, and honest uncertainty are more defensible than slogans on either side of the debate.