Research · Permitting & Environmental Review
The national data-center lesson: review the real project, not the label.
The most important permitting problem in the U.S. data-center boom is not that the country lacks review tools. It is that many hyperscale projects are outgrowing the labels used to approve them.
The takeaway
The public-policy response to data centers is often described as a backlash. That is partly true, but incomplete. Much of the response is a demand for decision-grade disclosure before approvals become hard to unwind. The national message is not 'stop building data centers.' It is 'stop approving undefined infrastructure campuses under soft labels.'
01
Review the real project, not the label
Across the country, large data-center campuses are being routed through local zoning, state environmental review, utility service planning, water and wastewater approvals, air permitting, wetlands review, and public-records processes that were not built for hundreds of acres of continuous-load industrial infrastructure. The failure mode is now familiar: the public file says 'technology center,' 'business park,' 'warehouse,' 'industrial campus,' or 'economic-development project,' while the real project is a power-intensive, water-relevant, generator-backed, utility-dependent hyperscale campus with regional consequences.
That mismatch is where the crackdown is coming from. Communities, utilities, regulators, and courts are not reacting only to data centers as a land use. They are reacting to under-defined data centers: projects whose electricity demand, backup generation, water use, wastewater profile, construction footprint, tax treatment, and public-infrastructure needs become clear only after approvals have momentum.
Project Skyway in Pine Island, Minnesota is a useful scenario inside that national pattern. It is not proof that Minnesota is closed to large data centers, and it is not a final appellate holding that Minnesota's Alternative Urban Areawide Review process is unlawful for data centers. The case is narrower and more transferable: when a public review record no longer matches the functional project being advanced, the review tool itself can become a schedule trap.
The practical rule for hyperscale sponsors and public agencies is simple: the public record does not need every commercial secret, but it must describe the environmental reality.
02
Data-center review is moving from land-use label to infrastructure reality
01
Local approvals are being asked to carry regional impacts
Data-center approvals often begin as local land-use decisions, but the consequences are regional: transmission upgrades, generation needs, water planning, wastewater treatment, construction traffic, tax-base effects, and ratepayer exposure. Virginia's JLARC framed this directly in its 2024 report: data-center growth creates statewide power-demand, cost, land-use, water, and air-quality questions that are not fully contained within a single locality's zoning vote.
02
Power demand has become an environmental and political fact
DOE and Lawrence Berkeley National Laboratory's 2024 United States Data Center Energy Usage Report projected rapid growth through 2028. A 24/7 hyperscale load is not just a customer account, it can require substations, transformers, feeders, transmission upgrades, generation commitments, construction power, and sometimes interim generation. If those pieces are deferred to later utility design, the public record looks incomplete because the project's operating premise has not been bounded.
03
Backup power is no longer a late-stage engineering detail
Diesel generators, gas turbines, fuel storage, testing cycles, emergency-use assumptions, demand-response participation, air emissions, noise, spill controls, and delivery logistics can drive environmental review and community opposition. A project that estimates greenhouse-gas emissions while leaving generator count, output, fuel, and operations to later design invites the obvious question: how was the impact estimate built?
04
Water risk is broader than cooling choice
Air cooling can reduce operating-water consumption compared with evaporative cooling, but it does not resolve peak-day demand, source-water capacity, drought resilience, wastewater routing, treatment capacity, construction dewatering, stormwater, floodplain, or wetlands impacts. In water-constrained or infrastructure-constrained communities, annual demand alone is not enough.
05
Public-records and confidentiality rules are now schedule issues
Hyperscale projects often involve NDAs, tenant secrecy, security-sensitive design, and competitive site selection. Those realities are legitimate. But when confidentiality is perceived to cover public-impact facts, use, acreage, load, water, wastewater, backup generation, mitigation, or public infrastructure, records disputes become permitting disputes. A public agency's document protocol can become as important to schedule as its engineering review.
06
The response is tightening
Local governments are adopting or considering data-center zoning limits, design standards, setbacks, noise rules, generator controls, water-disclosure requirements, and moratoria. State policymakers are examining ratepayer protections, tax-incentive conditions, utility large-load tariffs, and broader oversight. Environmental advocates are focusing on cumulative impacts, air emissions, water supply, public-records compliance, and whether projects were segmented or mislabeled. Utilities are increasingly asking for load certainty, financial security, minimum-demand commitments, and upgrade-cost allocation before promising service.
03
The Minnesota case that makes the national issue concrete
The case is Minnesota Center for Environmental Advocacy v. City of Pine Island; Ryan Companies US, Inc., Goodhue County District Court File No. 25-CV-25-2298. MCEA filed its complaint on October 16, 2025. The court first denied temporary relief on December 29, 2025 because construction was not expected before July 2026 and irreparable harm had not yet been shown. On May 22, 2026, after additional facts and a Minnesota Government Data Practices Act dispute had developed, the court granted a temporary restraining order, enjoined Ryan Companies from beginning or continuing construction or pre-construction activities within the Project Skyway AUAR study area, required a $2,000 bond, and denied the City/Ryan summary-judgment motion.
Those are interim litigation rulings. They are not a final appellate holding that the AUAR was unlawful. That distinction matters.
The Project Skyway Final AUAR was not silent about data centers. It evaluated a mixed light-industrial / technology-center Scenario A and a technology-center Scenario B. Scenario B covered approximately 440 acres of technology-center use, including a 3,000,000-square-foot technology center and 150,000 square feet of office space, and disclosed meaningful environmental quantities. Those disclosures cut against the easy narrative that 'nothing was reviewed.' The stronger conclusion is that the AUAR disclosed a concept-level technology-center envelope but did not fully bind or explain the project-specific assumptions that matter for a single-client hyperscale campus.
The May 22 order explains why that mattered. The court cited evidence that, before and during AUAR preparation, Ryan Companies and Pine Island had information pointing to a single-client project: a November 2024 nondisclosure agreement involving city administration and Ryan Companies; a similar agreement with a project consultant; a February 2025 'Skyway Community Talking Points' document describing the campus as being built 'for a single client, a US founded and headquartered Fortune 200 Company'; and alleged client involvement in draft-AUAR materials. The court did not finally decide liability. It found that the plaintiff had shown possible success on the merits and that genuine issues of material fact precluded summary judgment.
That is the national risk in miniature. A broad areawide envelope can be appropriate when a project is still an areawide planning scenario. It becomes fragile when opponents can plausibly show the sponsor and local government were no longer planning a generic market-driven envelope, but a specific hyperscale project with specific public-impact attributes.
Disclosed in the record
04
Seven recurring failure modes, not unique to AUARs, and not unique to Minnesota
01
The label is narrower than the infrastructure
A locality may approve a 'technology campus,' 'warehouse,' 'office/industrial park,' or 'data processing use' without a public record that shows the full infrastructure program: data halls, mechanical yards, electrical yards, generator yards, substation interface, utility corridors, water and sewer extensions, road improvements, security perimeter, stormwater basins, and construction staging. The fix is not semantic. It is a functional description: if the project is a hyperscale data-center campus, the record should say so and define the impact envelope.
02
The power case is treated as utility detail rather than public impact
Project Skyway had plausible transmission context, the Final AUAR noted a nearby Xcel Energy project to upgrade an existing 161 kV line and install a new 345 kV Mankato-to-Mississippi River line. That proximity makes service plausible. It does not prove deliverability. A credible review should disclose a bounded MW range by phase, the utility-service path, candidate service points, substation and transformer needs, transmission-upgrade dependencies, energization sequence, construction-power needs, and any interim or bridge-power strategy. Clean-energy procurement should be kept separate from physical deliverability.
03
Backup generation is left undefined
Generator count, fuel type, operating assumptions, emergency-use profile, testing profile, fuel storage, spill controls, air emissions, noise, and delivery logistics can affect environmental review, air permitting, stormwater risk, public safety, and community trust. The Project Skyway complaint alleged that the AUAR did not disclose generator number, output, or fuel source while still estimating GHG emissions. As a national siting lesson, the point is obvious: the backup-power envelope should be bounded before approvals harden.
04
Water is reduced to an annual number
Skyway's water profile is more nuanced than public controversy often allows, the AUAR stated air cooling was anticipated and that water cooling was not under consideration. But a national review standard needs more than annual demand. It should separate annual, maximum-month, peak-day, and phase-specific water demand; identify the source and appropriation basis; explain the relationship to current and projected municipal demand; describe drought and contingency operations; and separate domestic wastewater from process streams, cooling blowdown, treatment reject, washdown, or construction dewatering.
05
Stormwater, wetlands, floodplain, and geology are treated as later design
The AUAR identified Dry Run Creek, northern 100-year floodplain conditions, potential Wetlands C/D impacts of roughly 0.47 acre, and karst-prone geology with sinkholes within one mile. Those are serious but not necessarily fatal constraints. The broader lesson is design discipline: a large data-center campus should publish wetland delineation, avoidance/minimization sequence, floodplain and floodway findings, compensatory-storage assumptions, karst investigation plan, stormwater rate and water-quality controls, and any limits on infiltration caused by karst.
06
Construction impacts are understated because operations get the attention
A greenfield hyperscale campus is environmentally material before operations begin. Site preparation, grading, access roads, transformer and generator delivery, utility corridors, laydown, worker parking, truck peaks, lighting, noise, dust, erosion control, and adjacent receptors are part of the impact story. The review record should show how construction actually happens, not merely that construction will occur in phases.
07
Records compliance becomes schedule control
The records issue in Project Skyway was not a public-relations side story. By May 2026, MCEA asserted hyperlinks in produced emails were inaccessible, and the court found the Minnesota Government Data Practices Act request 'very much relevant' to the environmental-procedure lawsuit. The court found, for TRO purposes, that the City had not shown it complied with Minn. Stat. §13.03. It imposed only the $2,000 minimum bond because the TRO was at least partly due to the City's records response. NDAs may protect commercial or security-sensitive details. They should not be used, or appear to be used, to hide public-impact facts.
05
How to approve real hyperscale projects without building a litigation trap
01
Start with the highest known functional specificity
If the project is a hyperscale data-center campus, call it a hyperscale data-center campus. If it is single-client, multi-client, speculative, phased, or partially committed, say that in bounded terms. Tenant identity can be withheld where law permits. The use, scale, and impact envelope should not be vague. A useful public description includes site acreage, developable acreage, max and phase-specific building square footage, number and type of data-center buildings, mechanical and electrical yards, generator yards and fuel systems, substation and utility-interface areas, water/sewer/stormwater/road extensions, construction phases, adjacent receptors and buffers, and reduced-build alternatives.
02
Build the threshold matrix before choosing the pathway
Before relying on AUAR, EAW, categorical local review, site-plan approval, or any analogous pathway, practitioners should test the real project against mandatory review categories and permit triggers. In Minnesota, that means testing the actual envelope against Minn. R. 4410.3610, 4410.4300, and 4410.4400, plus water appropriation, air permitting, wetlands, public waters, wastewater, stormwater, energy infrastructure, and cumulative significance. The matrix should be public enough that readers can see why the pathway was selected.
03
Treat power as an environmental chapter
A large-load data center should not leave MW demand, service path, backup generation, bridge power, substations, transmission upgrades, and energization schedule to later utility detail. The public record should distinguish requested load versus studied load; initial phase versus full-build demand; physical service path versus clean-energy procurement; utility upgrades serving the project versus independent grid upgrades; temporary construction power versus permanent service; backup generation versus bridge generation; and customer-funded improvements versus rate-base or shared-network improvements.
04
Treat water, wastewater, and stormwater as separate issues
Air cooling can reduce operating-water risk, but it does not answer wastewater, peak demand, municipal capacity, drought, floodplain, karst, and stormwater questions. The minimum water package should include annual, peak-day, and phase-specific demand; source-water and appropriation basis; municipal-capacity margin; drought and contingency plan; wastewater character and route; treatment-capacity triggers; construction dewatering; stormwater controls; floodplain findings; and wetland avoidance/minimization.
05
Bound backup generation before approvals harden
The generator envelope should be disclosed early enough to inform air, noise, fuel, traffic, spill, and public-safety review. Where final equipment is not selected, the record can use conservative ranges. What should not happen is a final local approval based on an undefined generator fleet. The record should state generator count range, MW capacity range, fuel type, tank capacity range, testing hours, emergency assumptions, emissions basis, noise basis, delivery assumptions, spill controls, and whether generators could be used for non-emergency grid support or demand response.
06
Convert mitigation from concepts into commitments
Mitigation language should not read like a menu of things that 'may' be considered later. It should identify what will be imposed, who imposes it, when it is triggered, what approval makes it enforceable, what monitoring is required, what reporting is public, and what happens if the commitment is missed. For AUAR users this is especially important because the mitigation plan is the bridge between areawide review and later project implementation.
07
Build a records protocol before controversy
Hyperscale confidentiality is real. So is public-records law. The safest posture is to publish a clear log of what exists, what has been produced, what is withheld, and the legal basis for withholding. A strong protocol includes a public project docket, document index, production dates, redaction rules, statutory withholding bases, hyperlink and attachment capture, consultant-file handling, meeting-packet archiving, NDA status explanation, and a plain-language distinction between commercial secrets and public-impact facts.
08
Sequence entitlements around the environmental record
The more approvals advance while the environmental record is under attack, the more likely a court is to view temporary relief as necessary to preserve the status quo. If litigation or serious procedural objections arise before construction, the sponsor and agency should treat the period before mobilization as a cure window. A denied TRO is not necessarily a merits victory. If temporary relief is denied because construction is not imminent, the sponsor and agency should use the window to remove the mismatch that could support a renewed motion.
09
Give communities the impact file before they have to infer it
Opposition is not eliminated by better disclosure. But under-disclosure changes the character of opposition. It turns neighbors, advocates, utilities, and agencies into investigators. The more the public must infer the real project from NDAs, snippets, news reports, utility filings, or later disclosures, the more likely the process will migrate from planning rooms to courtrooms. The better model is a public impact file: project definition, load envelope, water package, generator package, construction package, mitigation commitments, records protocol, and response-to-comments matrix.
10
Plan for litigation as a schedule risk, not an afterthought
For large data-center campuses, litigation risk should be managed like power, water, and site-prep risk. The project schedule should include time for public comments, record correction, agency consultation, supplemental review, settlement, remand, or narrowed relief. The most expensive environmental document is often the one that has to be repaired after construction-readiness dates are already public.
06
Minnesota's AUAR process remains useful when the scenario is candid, bounded, and enforceable. The lesson is precise.
Define the project at the highest known functional specificity
If the project is a hyperscale data center, the review record should say so in environmental terms.
Build the mandatory-threshold matrix first
The pathway decision should be made against the real project envelope, including building area, energy infrastructure, air emissions, water, wastewater, wetlands, public waters, stormwater, and cumulative significance.
Do not rely on tenant confidentiality to obscure impact facts
Tenant identity and commercial terms may be protectable. Load, water, wastewater, backup generation, construction impacts, and mitigation are public-impact facts.
Make mitigation enforceable
AUAR mitigation should specify what will be imposed, who imposes it, when it is triggered, what approval makes it enforceable, and how compliance will be verified.
Treat records compliance as critical-path work
Public-records gaps can become irreparable-harm arguments when they prevent opponents from reviewing or using information before construction begins.
Use cure windows
If a court denies early temporary relief because construction is not imminent, the project should not treat that as a green light. It should treat it as time to fix the record.
In closing
Project Skyway's TRO was not inevitable. Opposition to a reported Google-linked hyperscale campus in Minnesota was foreseeable, but a construction-start injunction was not the only path. The avoidable risk was allowing the environmental-review record, records posture, and approval sequence to fall out of alignment with the project that opponents could plausibly show was being advanced.
The broader national lesson is bigger than AUAR. Data-center approvals are shifting from a land-use-label world to an infrastructure-reality world. Projects that disclose the real envelope early, power, water, wastewater, backup generation, construction, off-site infrastructure, public records, and enforceable mitigation, can still move. Projects that ask a generic label to carry a hyperscale reality are building a litigation schedule into the approval record.
Full brief · 13 pages · PDF
Read the complete brief
Standing note
Independent analysis based solely on publicly available court records, regulatory filings, environmental-review materials, and Environmental Quality Board materials. For educational and discussion purposes only. Does not constitute legal advice and does not represent any party to the Project Skyway litigation or development process.