Artificial intelligence (AI) data center development is accelerating across the country. New guidance from the U.S. Environmental Protection Agency (EPA) aims to help local leaders build these developments on designated Superfund or brownfield sites. While there are many factors that determine whether a community may decide to host an AI or “hyperscale” data center in general, the inclusion of Superfund or brownfield sites offers another option to revitalize these properties for economic development purposes.
On July 23, 2025, the White House released the Executive Order on Accelerating Federal Permitting of Data Center Infrastructure (EO) in alignment with the Administration’s goal to “facilitate the rapid and efficient buildout of this infrastructure by easing Federal regulatory burdens.”
The order contains directives to various agencies to support this effort, including providing financial support for qualifying projects, streamlining environmental and permitting and identifying suitable federal land for siting new data centers.
The EO directed EPA to “develop guidance to help expedite environmental reviews for qualified reuse and assist State governments and private parties to return such Brownfield Sites and Superfund Sites to productive use as expeditiously as possible.”
Pursuant to the EO, EPA issued new guidance (PDF) to help decision-makers identify sites that may be suitable for hosting AI data centers. The guidance includes specific considerations that must be evaluated in the development of AI data centers on Superfund and Brownfield sites.
Here are some key takeaways from the guidance for local leaders to understand.
What are Superfund and Brownfield Sites?
- Superfund sites include contaminants that pose risks to human health and the environment. They include manufacturing facilities, processing plants, landfills and mining sites that contain improperly managed hazardous substances and pollutants. Where there is no viable responsible party, EPA is granted authority and funding to clean up these sites.
- Brownfield sites are properties where current or future use is affected by concerns over environmental contamination due to commercial or industrial activities. EPA does not regulate brownfield sites but does provide “funding and technical assistance resources to support community-led efforts to revitalize and reuse” these sites.
Guidance on the Redevelopment of Superfund and Brownfield Sites as AI Data Centers
Prior to the EO, EPA already hosted a number of resources to assist local, state and tribal governments in revitalizing brownfield and superfund sites. Many of these are applicable in helping local leaders consider these sites for data center development, including:
- Leading Local Revitalization Fact Sheet (PDF)
- Navigating Funding for Brownfields Revitalizations Fact Sheet (PDF)
- Technical Assistance and Reuse Planning Support for Superfund Sites
- EPA’s Redevelopment Mapper for Superfund and Brownfield Sites
Key Considerations for AI Data Center Development on Superfund and Brownfield Sites
EPA has identified criteria for site characteristics that are compatible with AI data center development. The key things to consider are that AI data centers must be compatible with site conditions; have easy access to infrastructural support, including energy and fiberoptic cables; and be compatible with all applicable local, state and federal regulations.
Important Considerations when Siting AI Data Centers
Note: A full list of criteria can be found in Appendix A in the Guidance on the Redevelopment of Superfund and Brownfield Sites as AI Data Centers (PDF).
Size & Topography
- A minimum of 100 acres is typically needed for a 100 mega-watt (MW) campus-style data center. Data centers generating fewer MW may be accommodated on smaller sites.
- Additional land for power generation, substations, cooling, stormwater management, parking, security and phased expansion is usually desired.
- Rectangular sites with slopes less than 2-5 percent are generally preferred. 5-10 percent are riskier and slopes greater than 15-20 percent are not generally pursued.
Energy
- High demand for power requires sufficient grid capacity, substation access, redundancy and favorable tariffs. Onsite generation and backup sources are often needed to ensure uninterrupted power supply.
- Often, data centers are built directly next to or together with substations or power plants.
- Multiple power sources (hydro, nuclear and fossil) could be beneficial.
- Some data centers utilize renewable energy, such as solar, with solar arrays or turbines built on-site or located within a few miles.
Water
- Operations require substantial cooling, which may use millions of gallons per MW per year. Access to reclaimed or reliable water source is necessary.
- 1-10 miles from municipal water supply (with dedicated pipelines and large contractual agreements).
- Some data centers are directly adjacent to or less than 1 mile from major rivers/lakes or treatment plants (especially true for hyperscale sites in water-rich areas).
Fiber Optic Cable/Connectivity
- Redundant high-capacity connections via diverse routes required.
- Proximity to major internet exchanges, fiber optic networks and telecommunications infrastructure (often less than 1 mile) ensures low latency and high bandwidth. Cities with robust telecom infrastructure are preferred.
Next Steps for Local Leaders
- Learn more about AI Data Centers in NLC’s fact sheet series.
- Watch NLC’s webinars cohosted with the American Association for the Advancement of Science on data center water and grid impacts.
About the Author
What Does “Efficiency” Really Mean for Cities Right Now?
- NLC Partnerships
NLC’s Federal Update: DHS Shutdown, FEMA Review Council Extension and BRIC Funding
- Yucel Ors
AI in Local Government: Lessons from 2025 and Emerging Cyber Threats in 2026
- NLC Partnerships
Cybersecurity Isn’t the Problem: Cyber Risk Governance Is
- CyberAlliance
HelloNation: Bring Your City’s Story to Life
- CGI Digital
