Amazon says it has stopped using nondisclosure agreements with the government agencies it works with on data-center projects, a small wording change that could matter a lot for how AI infrastructure is approved. The statement appears in Amazon’s own data-center blog post, and TechCrunch reported that AWS chief executive Matt Garman used it while answering criticism that data-center deals are often negotiated behind closed doors.

The verified development is narrow. Amazon is publicly saying it no longer uses NDAs with government agencies on these projects. The retrieved material does not show when the policy changed, and it does not independently confirm the company’s claim. It also does not support the broader phrase “local governments” directly, so the safer reading is government agencies, as Amazon described them.

Why this matters beyond one line in a blog post

NDAs are not just paperwork. In practice, they can delay when public officials talk openly about a project, which in turn can shape when residents learn that a new data center is being planned. That is why the Amazon statement sits at the center of a broader trust problem around data centers: if disclosure comes late, the community debate starts after site selection, utility discussions, or permitting momentum are already well advanced.

Analytically, the practical consequence of dropping NDAs is earlier visibility, not automatic agreement. Communities may still oppose a project, but they are better positioned to ask about power demand, water use, backup generators, road impacts, tax arrangements, and grid upgrades before the deal hardens. That matters because the costs and benefits of a data center are often distributed unevenly: the operator gets capacity, while the local area absorbs traffic, land-use change and utility pressure unless the project is structured carefully.

That timing issue is especially sensitive because Amazon is arguing that data centers are now part of the country’s AI infrastructure buildout. In its own post, the company says AI is accelerating construction, that private industry is financing the expansion, and that more than 100 data-center moratoriums are being considered across the U.S. Those are Amazon’s claims, not independent findings in the retrieved material, but they explain why the company is trying to change the public narrative as well as the procurement process.

Amazon’s broader defense, in Amazon’s own numbers

The company’s blog post pairs the NDA statement with a longer defense of data centers. Amazon says direct data-center water use is about 0.5% of U.S. direct industrial water use, and it says its average data center uses less than 13,000 gallons per day. Amazon also says backup generators are idle 99.9% of the time, running roughly 10 hours a year for maintenance testing. Those figures come from Amazon’s post and should be read as company assertions rather than independently verified measurements.

Amazon also argues that communities receive tangible upside. In the same post, the company says it has contributed more than $1 billion to U.S. communities over the past three years in places where it has a meaningful data-center presence. It says projects can support schools, fire departments, police, roads and other local services through tax revenue and direct spending. Again, those are Amazon’s figures and claims, not external verification.

That is the real tension in the story. Amazon is making a case that data centers are critical infrastructure and that community benefits outweigh the disruption. Critics, meanwhile, say transparency is the first missing ingredient. If a company is serious about winning local consent for AI infrastructure, the strongest signal is not a new slogan about responsible buildouts; it is whether the company is willing to let public agencies speak before the approval path is locked in.

There is also a limitation that the story does not resolve: the retrieved material shows no broad federal or state reporting standard for data-center water, energy or community-impact disclosures. Without that baseline, Amazon’s NDA pledge may improve openness in some cases, but it does not by itself make the underlying data easy to compare across projects or operators. That is why the policy change is significant, but not conclusive.

Sources: Amazon; TechCrunch.

If you follow a data-center proposal, check whether permit filings and community briefings appear before approvals, because that is where Amazon’s NDA pledge becomes meaningful transparency.