OpenAI promises to pay for AI data center electricity as voters push back against rising power bills

OpenAI is promising to pay the full electricity costs of its future AI data centers, including the infrastructure needed to run them, just two days after American voters delivered a sweeping rejection of data center expansion in the 2026 midterm elections.

The ChatGPT maker announced a new set of commitments Thursday to address growing concerns over electricity bills, water consumption, environmental damage, and the enormous computing facilities being built across the United States.

Under its new National Community Compact, OpenAI pledged to cover the full cost of electricity and related infrastructure at future data centers it owns and operates. The company said it would provide financial guarantees to protect utility customers if projects are delayed, scaled back, or abandoned, CNBC reported.

The announcement comes at a politically sensitive moment for an AI industry spending hundreds of billions of dollars on computing infrastructure. Across the country, communities are increasingly questioning whether the economic benefits of these massive facilities justify their demands on local resources.

Voters’ message this week was hard to miss.

Of 41 local ballot measures across 10 states addressing data center development, voters supported restrictions or rejected measures favoring development in all but one, according to an analysis by Ballotpedia and election results compiled by CNBC.

The results suggest that opposition to AI infrastructure has moved beyond local planning meetings and become a political issue with consequences for companies racing to build the next generation of computing facilities.

OpenAI wants to reassure communities that AI growth won’t raise their electricity bills

OpenAI’s latest commitments extend beyond electricity costs.

The company said it would seek community feedback before construction, disclose development plans, publish information about water and electricity consumption, and commission independent noise studies.

It pledged to avoid nondisclosure agreements with public officials, apart from limited protections for sensitive information, giving residents greater visibility into negotiations that have sometimes taken place behind closed doors.

In a separate Compute Campus Policy Blueprint, OpenAI urged federal, state, and local lawmakers to turn similar commitments into enforceable industry standards.

“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,” OpenAI Chief Global Affairs Officer Chris Lehane told reporters during a Wednesday briefing. “This will help ensure that there is a very high floor in terms of how people move forward.”

OpenAI has already begun putting some of these ideas into practice.

In Georgia, the company has committed to covering electricity and infrastructure costs associated with its planned Effingham County data center. Its agreement with Georgia Power includes provisions allowing electricity deliveries to the campus to be reduced by up to 1,000 megawatts during periods of high demand.

That arrangement is intended to reduce pressure on the grid when households and businesses need electricity most.

OpenAI has made similar community commitments in Ohio, where it is pursuing major computing infrastructure projects as part of its broader expansion plans.

The financial stakes are enormous. OpenAI reportedly told investors earlier this year that it was targeting approximately $600 billion in cumulative computing spending by 2030, reflecting the extraordinary cost of building and operating the infrastructure behind advanced AI systems.

Yet the political environment surrounding those investments is changing.

For residents worried about household expenses, promises of new jobs and economic development may carry less weight than concerns about electricity bills, water supplies, and the long-term effects of industrial-scale computing facilities.

OpenAI’s new commitments acknowledge those concerns, but voluntary promises are different from legally enforceable protections. Their effectiveness will depend on how the company structures utility agreements, reports costs, and follows through on its obligations.

The larger question now extends beyond OpenAI.

As AI companies compete to build bigger data centers, voters are making it increasingly clear that communities expect a say in how those facilities are built and who pays for them.

For an industry accustomed to moving quickly, winning public support could become just as important as securing the electricity needed to keep its servers running.