Over the summer, the White House released its “AI Action Plan,” setting forth a host of policy pronouncements for the advancement and support of U.S. dominance in AI. One of the policy “pillars” should have all environmental lawyers sit up and take notice – support for building AI infrastructure. An entire “Data Center Infrastructure” Executive Order accompanied the Action Plan, and outlined a series of regulatory plans intended to facilitate data center construction and related energy infrastructure, including:
- Having CEQ establish new categorical exclusions to speed NEPA review;
- Directing EPA to expedite federal permitting and to identify Brownfield and Superfund sites for project siting;
- Calling for programmatic Endangered Species Act consultation from Departments of Interior and Commerce;
- Requiring Army Corps review of Section 404 nationwide permits (and develop a new one as needed) for projects; and
- Promoting use of federal lands under control of the Interior, Energy, and Defense Departments.
EPA has already acted in response to this EO, announcing in September that it will prioritize TSCA review for “new chemicals related to data centers” and AI projects, departing from the Biden Administration’ priorities under TSCA for expedited reviews for chemicals utilized in renewable energy development and battery components. More regulatory activity is sure to follow but will likely need to await the end of the current government shutdown.
The broad reach of environmental regulatory and compliance activity is unsurprising because the brick and mortar needs for AI in data centers require land, water, and energy at a scale not seen before in the communities where they now reside (such as the hundreds of facilities in “Data Center Alley” in northern Virginia). But data centers also present opportunity for communities looking for beneficial reuse options for former industrial property from the last century that may be well situated both near water supplies and connection to the grid, and on land that does not intersect threatened or endangered wildlife habitat. Land that went through the birth of environmental regulation and permitting may be the best candidate to do so again to power the AI revolution.
However, data center developers who disregard environmental compliance do so at their peril, as Elon Musk’s xAI learned this summer, receiving a Clean Air Act citizen suit Notice of Intent to Sue from the installation and use of unpermitted gas turbines at a data center in Memphis, Tennessee. In response, and presumably in the hopes of avoiding citizen suit litigation through a diligent prosecution defense, xAI received an air permit, which is now being appealed by local stakeholders.
What is clear is that 21st century AI firms and data center developers are going to have to navigate an environmental legal landscape that was very familiar to 20th century industry – land zoning, water use, energy procurement and consumption, permitting of emissions and releases, resource consultations, compliance and auditing, and community relations and engagement. While it appears they may benefit from some regulatory streamlining and efficiencies if federal agencies execute the ambitious AI Action Plan, they will not escape our modern day environmental regulatory regimes (including attendant litigation and enforcement). We should therefore expect environmental lawyers to have an important role to play in this 21st century industrial tech revolution.
