[OpenAI](https://openai.com) just threw open the curtain on data center dealmaking. The company announced today it’s ending the practice of requiring non-disclosure agreements with public officials when discussing data center projects, a reversal that comes as communities nationwide revolt against AI infrastructure they say is being forced on them in secret.
The timing isn’t subtle. Just days after [midterm elections saw voters reject data center expansion](https://www.theverge.com/policy/1026365/data-centers-midterms-results) in multiple jurisdictions, OpenAI published [new commitments](https://openai.com/index/setting-new-standard-responsible-ai-infrastructure/) promising it will only seek “narrow protections covering legitimate security information, trade secrets, or proprietary or competitively sensitive information.” Translation: no more blanket gag orders on elected officials trying to discuss what’s coming to their towns.
Data center NDAs have become a flashpoint as AI companies race to build the massive computing facilities needed to train and run large language models. Critics say the agreements let developers and local officials cut deals behind closed doors, leaving residents in the dark about projects that can strain power grids, drain water resources, and industrialize quiet communities. [NPR reported](https://www.npr.org/2026/08/27/nx-s1-5879528/data-center-nda-disclosure-louisiana) how these arrangements have hidden crucial discussions in Louisiana and other states.
[OpenAI’s](https://openai.com) move follows [Amazon’s October announcement](https://www.theverge.com/tech/1003929/amazon-ai-data-center-blog-warning) that it would stop using NDAs with public officials on data center projects. That commitment came with its own warning shot, as Amazon cautioned communities that rejecting its facilities could mean missing out on economic benefits. The message was clear: we’ll be more transparent, but don’t expect us to take no for an answer.
The ChatGPT maker is trying to get ahead of a building wave of opposition. [Earlier this week, OpenAI pledged to cover electricity costs](https://www.techbuzz.ai/articles/openai-pledges-to-pay-data-center-power-costs-after-voter-revolt) for communities hosting its data centers and promised to involve residents in planning decisions. That announcement came directly after midterm results showed voters aren’t buying the jobs-and-growth pitch that typically accompanies these projects.
But OpenAI’s transparency offensive goes beyond NDAs. The company says it’s committing to community input processes, ratepayer protection, and working with local utilities to ensure grid stability. These aren’t just nice-to-haves anymore – they’re survival tactics for an industry facing organized resistance.
The backlash reflects a broader reckoning over AI’s physical footprint. Training models like GPT-4 requires enormous amounts of electricity, and running them at scale demands data centers that can consume as much power as small cities. [Microsoft](https://microsoft.com), [Google](https://google.com), and [Meta](https://meta.com) are all racing to build similar facilities, often in areas with cheap power and favorable tax treatment.
What’s changed is that communities are wising up to the tradeoffs. A data center might bring construction jobs and property tax revenue, but it also brings 24/7 noise, increased truck traffic, pressure on water supplies for cooling, and electricity demand that can drive up rates for everyone else. When those discussions happen under NDA, residents only find out when it’s too late to object.
[OpenAI’s](https://openai.com) policy shift suggests the company sees where this is heading. With [CEO Sam Altman courting Gulf states for AI infrastructure deals](https://www.techbuzz.ai/articles/openai-inks-four-abu-dhabi-deals-as-altman-courts-gulf-ai-power) and the company’s computing needs growing exponentially, OpenAI can’t afford to have every data center project turn into a referendum on AI itself.
The question now is whether transparency alone will satisfy communities or if voters want veto power over projects that reshape their energy landscape. OpenAI’s betting that sunlight is enough disinfectant. The midterm results suggest residents might want more than promises – they want control.
OpenAI’s NDA reversal marks a turning point for AI infrastructure politics, but transparency is just the opening bid. As training runs get bigger and inference demands explode, every major AI company will need social license to build at scale. The midterms proved that license can’t be bought with tax breaks or negotiated behind closed doors anymore. Communities want a seat at the table before the concrete gets poured, and they’re willing to vote down projects that ignore them. For OpenAI and its rivals, that means the next bottleneck for AI progress might not be chips or algorithms – it could be town hall meetings.
