Amazon Ditches NDAs for Data Center Deals as Community Backlash Grows

Amazon says it will stop using NDAs when negotiating data center deals with local governments. The company announced the shift after Microsoft made a similar move earlier this year, and the timing is not incidental: secrecy around data center siting has fueled community backlash against AI infrastructure across the United States.

The scale of that backlash is measurable. Opposition to data center construction has produced hundreds of proposed and enacted moratoriums, stretching from New York to San Francisco. Local governments and residents have pushed back against projects whose terms they could not see, and the nondisclosure agreements that Amazon and other hyperscalers routinely demanded during negotiations became a flashpoint in that fight.

The story was the central topic on a recent episode of TechCrunch's Equity podcast, where hosts Anthony Ha, Sean O'Kane, and Rebecca Bellan examined whether dropping NDAs will actually change how data center deals get done.

Why are data center NDAs under fire?

Nondisclosure agreements have long been standard practice when cloud providers negotiate with municipalities over land, power, water, and tax incentives for data center campuses. The deals are large, competitive, and often involve concessions from local governments that residents only learn about after the fact.

That opacity has collided with a building boom driven by AI compute demand. As the Equity hosts discussed, communities from New York to San Francisco have responded with moratoriums — hundreds of them, proposed and enacted — that pause or block new data center construction.

Amazon's decision to stop using NDAs in these negotiations, following Microsoft's earlier this year, signals that at least some major operators now calculate that transparency costs less than the permitting delays and political resistance that secrecy invites. Whether the change is cosmetic or structural remains the open question the podcast hosts pressed on.

What does transparency actually change?

The Equity discussion raised a straightforward test: does ending NDA practice alter the substance of how deals are struck, or only the optics around them?

Two facts anchor the debate:

• Amazon has committed to dropping NDAs in negotiations with local governments for data center deals.
• Microsoft made the same move earlier this year, making Amazon the second major cloud operator to do so.

The hosts also connected the trend to the broader political environment. Hundreds of moratoriums across jurisdictions from New York to San Francisco represent organized, local-level resistance — not isolated complaints. For companies spending billions on AI infrastructure, each moratorium translates into delayed capacity in a market where compute supply is the binding constraint.

Secrecy, in other words, has a price. The industry's turn toward disclosure suggests that price has finally exceeded the competitive advantage of keeping negotiation terms hidden from the communities being asked to host these facilities.

Is trust the real product for personal AI startups?

The podcast's second thread moved from infrastructure to consumer software. A wave of startups is now betting that consumers will hand AI agents access to their inboxes, files, and credit cards — but only if those agents can get websites to let them in.

That framing carries a striking parallel to the data center story. In both cases, the barrier is not raw capability but permission: AI systems need access — to municipal negotiation rooms in one case, to personal data and commercial websites in the other — and access depends on trust that has not yet been earned.

As the Equity hosts discussed, trust may be the real product for these personal AI startups. The companies in this space are not merely selling automation; they are asking users to delegate sensitive capabilities, from reading email to making purchases, to software that acts on their behalf.

The gating problem is external as well as internal. These agents must also convince websites to allow them in, a gatekeeping layer that sits between the startup's product and the user's intent. Without that cooperation, consumer AI agents remain boxed out of the very services they are designed to operate.

What comes next?

Both storylines converge on a single question for the AI industry: can it operate in the open? Amazon and Microsoft dropping NDAs tests whether transparency at the infrastructure layer reduces community opposition or simply exposes deal terms to fresh scrutiny. The personal agent startups test whether consumers will extend to AI the access they already give to banks, email providers, and payment networks.

The Equity episode closed with the hosts' regular deals-of-the-week segment, continuing its coverage of funding and M&A across the startup ecosystem.

The stakes in both cases are concrete. On the infrastructure side, hundreds of moratoriums from New York to San Francisco already constrain where AI data centers can be built, and every negotiation conducted under new transparency rules will test whether disclosure speeds or stalls the next wave of projects. On the consumer side, the startups betting on inbox, file, and credit card access are wagering their businesses on a proposition that remains unproven: that users will trust AI agents enough to open the door — if websites open it first.