OpenAI released a National Community Compact and policy blueprint on 5 November 2026, pledging that future OpenAI data centres it owns and operates will pay their own electricity and grid costs while promising not to require public officials to sign non-disclosure agreements. The framework urges US regulators to turn the voluntary pledges into binding industry rules.
Key points
- OpenAI promised that future campuses it owns and operates will pay their full utility expenses and provide financial guarantees if a project is delayed, reduced or cancelled.
- The company will not require public officials to sign non-disclosure agreements and will disclose site-level water and electricity consumption.
- A companion policy blueprint urges federal and state regulators to establish large-load power tariffs and tax incentive clawbacks across the AI sector.
- The announcement follows US elections on 3 November 2026, where voters backed restrictions or rejected pro-development proposals in 40 out of 41 local ballot measures across 10 states.
Pledges for future campuses
The National Community Compact establishes operating standards for every future AI campus that OpenAI owns and operates. Under the energy terms, the company committed to covering the cost of the electricity it uses and the physical grid infrastructure needed to supply its facilities.
To shield other utility customers, OpenAI will provide financial assurances before major capital expenditure begins. These protections ensure that ratepayers do not bear the cost if a facility is delayed, scaled back, or cancelled. The company also promised to supply electric utilities with realistic projections of its power demand and schedule to assist with regional transmission planning.
Community engagement rules in the compact require OpenAI to consult residents before construction starts. The company will use listening sessions and public surveys, appoint a permanent local contact person for the life of each facility, and publish annual progress reports.
Secrecy rules and resource reporting
OpenAI said it will avoid non-disclosure agreements when negotiating data centre deals with public officials. The company will restrict confidentiality clauses to narrow protections for security measures, trade secrets, and proprietary operational information.
For construction, OpenAI plans to release site-specific mitigation strategies for noise, lighting, and heavy traffic. The company committed to hiring independent specialists to carry out noise assessments before operations begin and each year during operations.
Water consumption has drawn public scrutiny near AI developments. OpenAI pledged that each campus will adopt closed-loop or low-water cooling systems and publicly report site-level water and electricity usage. The company also said it will publish workforce strategies, create apprenticeship opportunities, and set local hiring targets where labour conditions permit.
Georgia and Ohio agreements
The pledges build on early initiatives OpenAI began testing earlier in 2026. In January, the company launched its Stargate Community framework, followed on 2 November by Grid Protectors, a scheme designed to curb facility power demand during periods of extreme electrical grid stress.
OpenAI has already linked several of these mechanisms to live construction projects. In Effingham County, Georgia, the company committed $80 million to local priorities. Those funds include a $4.8 million grant to provide free school meals for all 14,000 public school students in the county, alongside $2 million to build a municipal park.
The Georgia site, known as Project Camellia, involves a commercial contract with utility Georgia Power for 3.2 gigawatts of electricity, scheduled for phased delivery between 2028 and 2032. Under the contract, Georgia Power holds the right to reduce power flow to the facility by up to 1,000 megawatts when regional demand peaks.
In Pike County, Ohio, OpenAI partnered with clean-energy developer SB Energy to invest $6 million in local childcare facilities. The expansion fits into OpenAI’s broader infrastructure push, after the company told investors in February that it targets $600 billion in total compute expenditure by 2030, a programme detailed in its plans for dedicated AI compute campuses.
What regulations is OpenAI proposing for AI data centres?
The release came two days after American voters voiced direct opposition to AI infrastructure at the ballot box. On 3 November 2026, voters considered 41 local ballot measures across 10 states concerning data centre development, backing restrictions or rejecting pro-growth initiatives in 40 of them, according to data from Ballotpedia and CNBC.
Public resentment mounted over the summer as household expenses, grocery bills, and petrol prices rose, turning AI data centres into a major US midterm election flashpoint. OpenAI acknowledged in its blueprint that voluntary corporate pledges alone cannot give communities sufficient certainty as computing demands expand across the country.
Alongside its voluntary compact, OpenAI published its Compute Campus Policy Blueprint, calling on local, state, and federal lawmakers to turn its commitments into statutory requirements for all AI operators.
The document advises state public utility commissions to establish large-load electricity tariffs. These tariffs would mandate that high-demand industrial power consumers pay the specific grid expansion expenses they cause, accompanied by public accounting of cost allocations. The blueprint also backs demand-flexibility rules with clear operating conditions, including load curtailment, on-site battery storage, and local generation during supply shortages.
At the federal level, OpenAI urged lawmakers to streamline permitting for energy generation and transmission, and grant the Federal Energy Regulatory Commission explicit authority to accelerate grid interconnections for shovel-ready generation projects, subject to reliability and ratepayer protections. It also recommended reinstating and broadening technology-neutral clean-electricity tax incentives, attaching mandatory clawback clauses to discretionary tax subsidies if developers miss material commitments, and publishing early purchasing plans so local suppliers can compete for contracts.
Chris Lehane, OpenAI’s chief global affairs officer, told reporters during a briefing: “What we’re doing with this blueprint is really laying out the best lessons that we have learned and committing to applying them as they go forward.”
Lehane said the policy recommendations would create a standard for the industry: “This will help ensure that there is a very high floor in terms of how people move forward.”
OpenAI said it has held discussions on the blueprint with political candidates and lawmakers since August. Lehane said having those conversations now helps build momentum so that work is already done in advance when the new Congress convenes in late January.
