Voters rejected data centers in 40 of 41 ballot measures. OpenAI's answer: pay its own power bills
Two days after voters backed data center restrictions in 40 of 41 ballot measures across 10 states, OpenAI pledged to pay full power costs and pressed policymakers to make such rules law.

Why it matters
- Voters in 10 states backed data center restrictions or rejected pro-development measures in 40 of 41 ballot measures on Tuesday, per Ballotpedia and CNBC.
- OpenAI on Thursday pledged to pay the full cost of powering data centers it owns and operates, publish construction plans and avoid NDAs with public officials.
- OpenAI told investors in February it targets $600 billion in total compute spend by 2030, with major data center projects announced in Georgia and Ohio.
- Chief Global Affairs Officer Chris Lehane said the policy blueprint has been in the works since August and followed meetings with dozens of lawmakers.
Voters rejected or restricted data center development in 40 of 41 local ballot measures across 10 states on Tuesday — and two days later, OpenAI published a set of commitments pledging to pay the full cost of powering the data centers it owns and operates.
The commitments, released Thursday, go beyond the power pledge. OpenAI said it will solicit community input on its facilities, publish construction plans, and stay away from non-disclosure agreements with public officials. Alongside the commitments, the company issued a policy blueprint encouraging lawmakers to turn similar rules into law.
The stakes are straightforward. AI companies like OpenAI depend on data centers to serve customers and train new models, and they're racing to build capacity as fast as demand grows. In February, OpenAI told investors it's targeting $600 billion in total compute spend by 2030. The company has since announced massive data center projects in Georgia and Ohio. When local voters block those facilities, the buildout stalls.
What did OpenAI actually promise?
The core commitment is financial: OpenAI will cover the full cost of powering data centers it owns and operates. That matters because one of the loudest public criticisms of the industry is that data centers strain local grids and push utility bills up for residents who get nothing in return.
The other pledges target transparency:
- Solicit community input before building
- Publish construction plans
- Avoid NDAs with public officials
The last item speaks to a specific frustration. Residents and local journalists have complained that data center deals are often negotiated quietly, with officials bound by confidentiality agreements until the terms are locked in. OpenAI is now publicly ruling out that practice for its own facilities.
Crucially, these commitments apply only to data centers OpenAI owns and operates. Much of the industry's compute sits in facilities run by other providers, where OpenAI's pledges carry no weight.
Why did OpenAI publish this now?
The timing is the story. OpenAI released the commitments two days after voters delivered what CNBC calls a stinging rebuke of data centers at the polls. Out of 41 local ballot measures across 10 states, voters backed restrictions or rejected pro-development measures for all but one, according to an analysis by Ballotpedia and additional results gathered by CNBC on Wednesday.
That near-shutout reflects opposition that built quickly over the summer, and across party lines. Consumers already grappling with inflation, rising grocery prices and stubbornly high gas prices grew worried that data centers could worsen environmental harms, raise utility bills and contribute to job losses. For many voters, the facilities became a physical manifestation of the broader anxiety around AI itself.
The takeaway: local resistance to data centers has moved from scattered complaints to a proven ability to win at the ballot box, and OpenAI is responding before the trend hardens.
What does the policy blueprint propose?
OpenAI didn't stop at voluntary commitments. Its policy blueprint encourages policymakers to enact similar requirements as law — effectively asking governments to make the company's own rules binding on everyone.
Chris Lehane, OpenAI's Chief Global Affairs Officer, framed the blueprint as a distillation of what the company has learned through its buildout so far.
"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 told reporters during a Wednesday briefing. "This will help ensure that there is a very high floor in terms of how people move forward."
That "high floor" language signals the strategic logic. If every operator has to meet the same transparency and cost standards, OpenAI gives up little competitive ground — and the industry as a whole sheds its worst actors as political targets.
Who has OpenAI been talking to?
The blueprint has been in the works since August, Lehane said. In that time, he has met with dozens of lawmakers and candidates.
Lehane also said the commitments build on pledges OpenAI has already made this year to protect ratepayers and invest in local communities. The company has repeated those promises repeatedly through 2024 as public opposition mounted, and Thursday's announcement folds them into a single formal package.
The takeaway: this is a months-long lobbying effort reaching its public phase, not a rapid reaction to a single election night.
Is OpenAI aiming at Congress?
Yes, by Lehane's own account. He tied the blueprint's release to the legislative calendar, arguing that groundwork laid now pays off when the new Congress convenes.
"Having these conversations now actually helps put more energy into the momentum, so that when people come back in late January, new Congress is in in session, a lot of the work is being done in advance," Lehane said.
That comment reveals the play. OpenAI would rather shape federal or state rules it helped draft than face a patchwork of local ballot measures it can't control. The election results show exactly why: 40 out of 41 votes went against the industry's buildout model.
Why the money matters
The $600 billion compute-spend target through 2030 is the number that explains everything else in this story. OpenAI can't reach it without gigawatts of new data center capacity, and Georgia and Ohio projects are already announced. Each local ban or restriction chips away at that runway.
Voters, meanwhile, have shown they see data centers as an AI cost without a visible local benefit — higher bills, environmental risk, and potential job losses, set against infrastructure that serves customers elsewhere. OpenAI's pledge to pay full power costs is a direct attempt to break that equation.
The open question is whether voluntary commitments and a friendly blueprint can outpace the ballot box. Tuesday's results suggest voters aren't waiting for Washington — and every additional restrictive vote will land before any law Lehane helps write does.
For now, OpenAI has put its terms on the table: transparency, no NDAs with officials, community input, and paid power bills. Whether statehouses and the new Congress adopt them — and whether other operators follow — will decide if the industry can keep building at the pace its investors expect.
Source: CNBC Tech
More from Sophie Lindqvist
Show full bio
Staff writer covering regulation, policy and their impact on artificial intelligence and machine learning at AI In Context.
367 articles
Related articles
- OpenAI Pledges 3.2 GW Data Center in Georgia With $80M Community Fund
- OpenAI pitches its data centers as emergency power reserves
- Data Center Backlash Becomes a Political Force After Midterms
- OpenAI and Oracle Commit to 4.5 GW of New Stargate Data Centers
- NVIDIA to invest up to $100 billion in OpenAI for 10 gigawatts of AI infrastructure