OpenAI says future campuses it owns will skip NDAs with officials and pay their own power costs
The 5 November post publishes a National Community Compact and a policy blueprint. Each campus OpenAI owns and operates is supposed to get a local compact: full electricity and hookup costs, a water plan, annual reporting, and no broad NDAs with public officials. The Verge’s note, just after midnight UTC on Friday, ties the announcement to this week’s elections.
Feature3 minIn-depth
Eli Navarro3 min read
- OpenAI
- Data centers
- Energy
- Policy

Photo: OpenAISource page
OpenAI on 5 November published a National Community Compact and a Compute Campus Policy Blueprint for the data centers it is starting to call compute campuses. The compact is what the company says communities can expect at future sites that OpenAI itself owns and operates. Each of those sites, the post says, will have its own local compact, written around local priorities, with a shared list underneath: early engagement, a local point of contact for the life of the campus, and a public report every year. Jay Peters at The Verge, in a story timestamped 12:37 a.m. UTC on Friday, describes the announcement as a response to this week’s elections, in which he says voters opposed data-center expansion. That electoral claim is his frame. It is not in OpenAI’s post.
The energy line is the one OpenAI states as a cost, not a rate. It says it will pay the full cost of the electricity it uses and of the infrastructure needed to serve it. Before major spending begins, it says it will provide financial assurances so other customers are protected if a project is delayed, reduced, or canceled. On water, it says each campus gets a plan, that it will prioritize closed-loop or low-water cooling, and that it will publish site-level water and electricity use. Construction plans and independent noise studies are supposed to be public before operations and then once a year. On secrecy, the post says OpenAI will not use non-disclosure agreements with public officials, “allowing only narrow protections covering legitimate security information, trade secrets, or proprietary or competitively sensitive information.” Peters quotes that limitation and adds that Amazon said in October it would no longer use NDAs with government agencies on data-center projects. That Amazon comparison is in The Verge’s story.
Two counties, and a blueprint that asks for rules
The jobs commitment is a publication schedule and a hiring report, plus work with unions and training providers. OpenAI also says it will keep investing in priorities residents help choose. The examples it gives are already local. In Effingham County, Georgia, residents shaped a $4.8 million grant for free school meals for all 14,000 students in the county. In Pike County, Ohio, families named childcare, and OpenAI and SB Energy made a joint $6 million commitment. The post says this continues Stargate community commitments announced in January, and later commitments in Georgia and Ohio on energy costs, water, local investment, and accountability.
The blueprint is addressed to lawmakers, and OpenAI says voluntary company promises are only a start. It asks state utility commissions to require large electricity customers to cover the costs they add and to protect other customers if a project falls through. It says utilities should make it easier to connect campuses that can cut demand when power is short. At the planned Effingham campus, it says, the agreement with Georgia Power lets the utility reduce power delivered to OpenAI by up to 1,000 megawatts when demand is high. The post also asks Congress to speed transmission upgrades and to restore and expand tax incentives for clean electricity. Where governments offer tax breaks, OpenAI says the terms should be public, the commitments measurable, and the benefits recoverable if a developer falls short. Developers, in the blueprint’s account, should share purchasing plans early enough for local suppliers to bid, and should publish hiring forecasts that unions and community colleges can train against.


