Data Center Transparency: Can Ending NDAs Build Trust?
Learn why data center transparency matters, what ending NDAs can change, and how AI companies can earn community trust before projects are built in practice.
The AI Infrastructure Debate Is Really About Trust
AI companies want more data centers. Communities near proposed sites want real answers about power bills, water supplies, property values, and how local plans will change. Those concerns aren't something you can smooth over with a better press release. They're basic questions about costs, benefits, and who gets a say before decisions get locked in.
Amazon says it will stop using nondisclosure agreements (NDAs) while negotiating with local governments over data centers, following a similar move by Microsoft. That matters because confidential talks can leave residents reacting to a project after the important choices have already been made. Still, dropping an NDA doesn't automatically make a project transparent, beneficial, or acceptable. It's a first step - not a full trust plan.
The real question is what companies should disclose, when they should disclose it, and what they should do when a community says no. This piece looks at the limits of NDA-free negotiations, why temporary moratoriums still matter, and a practical framework for earning a hearing before an AI infrastructure project moves forward.
Why data center projects have become a trust issue
A data center is physical infrastructure. It needs land, ties into the electricity grid, requires cooling, and can shape local decisions about zoning, roads, and utilities. That makes it different from software. People don't "install" a data center or ignore it after it shows up - they live with what it does.
And even if it's part of an AI supply chain, the local effects are concrete: energy demand, water use, construction traffic, noise, and land-use changes. If communities start with those questions, that's reasonable.
People aren't automatically trying to block development. They're trying to evaluate a proposal:
- Will this add pressure to a constrained grid?
- How much water will be used, and where will it come from?
- Will local taxpayers pay for upgrades?
- How many jobs are permanent, and how many are just for construction?
A company can be technically correct and still be the wrong fit for a specific location. Likewise, a community can support a project when costs are clear, benefits are credible, and commitments are enforceable.
The information gap starts before the public meeting
When negotiations happen privately, residents may not know who is proposing the facility, what it's meant to support, or which public incentives are on the table. Sometimes officials also can't share much while talks are underway. By the time details are released, the project can feel like it's already settled.
That timing changes how people react. A late-stage public meeting can feel more like a presentation than a real chance to influence outcomes. Even accurate information may be met with skepticism if people believe the central decisions were made without them.
This pattern shows up in other tech-driven expansion. During the rapid growth of ride-hailing in the 2010s, some companies pursued local deals and subsidies through closed-door negotiations. The problem wasn't only secrecy - it was the concern that private bargaining could make it easier to ask for additional concessions without the public seeing what was being traded. The same risk exists when governments negotiate incentives or infrastructure arrangements with data center developers.
Communication is not a substitute for accountability
Yes, the AI industry has a communication problem. Companies often assume that people will "get it" because the technology looks impressive to insiders. But a resident worried about a utility bill doesn't need another abstract promise about an AI-powered future. They need a straight account of what the facility will consume, what it will deliver, and what happens if assumptions change.
Better explanations help - but they can't fix weak terms. If a company can't back up claims with project-specific data, it should say that. If most jobs are temporary, it should distinguish construction work from long-term positions. If water use depends on weather, cooling design, or operational choices, those conditions should be explained instead of buried behind a single headline number.
That's why the broader debate about AI ethics and accountability matters for infrastructure. Trust isn't something a company can ask for. It's what you get when decisions are visible enough to be questioned and tested.
What ending NDAs changes - and what it does not
Amazon's announcement, reported alongside its case for community benefits, is meaningful because it changes one part of the negotiation process. Microsoft made a similar commitment earlier in 2026. If these policies are applied consistently, local officials should have more room to discuss proposed projects publicly instead of being boxed in by confidentiality agreements.
That can improve the conversation. Residents, journalists, utility providers, and elected officials can look at the same baseline information while decisions are still open. Companies also have a chance to hear objections before they turn into organized opposition. None of this guarantees agreement, but it makes informed disagreement more possible.
An NDA-free approach also sends a signal: the company is willing to put at least some of the discussion behind a major local decision in the open. For an industry that often frames infrastructure as essential to economic progress, that's a fair minimum expectation.
Transparency needs a defined scope
Removing an NDA doesn't mean every document, security detail, or commercial term gets published. Some information may legitimately need protection. A company might have good reasons not to disclose sensitive network architecture, proprietary system designs, or anything that would create a security risk. Local governments also have to follow procurement rules and public-records requirements.
The practical question is narrower: is confidentiality being used for a limited, defensible purpose - or broadly enough that residents can't understand the parts of the deal that affect them?
A credible disclosure approach should make the boundaries clear:
- What information is being withheld, and why
- What project details are available publicly
- What incentives or infrastructure commitments are being considered, where disclosure is allowed
- Who will answer questions, and how residents can submit them
- When the next update will be published
If the company and local government can't explain those limits, it's reasonable for the public to assume the secrecy exists to protect the deal - not to protect the community's ability to evaluate it.
An announcement is not the same as an operating policy
Companies also need to define what "ending NDAs" actually covers. Does it apply only to negotiations with local governments, or does it extend to utilities and other public bodies? Does it cover all new projects? What about existing agreements? Can a municipality still be asked to keep basic project information confidential under some other contract clause?
Those details matter because a policy can sound stronger than it is. A useful commitment should be specific enough that journalists and residents can check whether it was followed. Otherwise, a company may drop one formal NDA while leaving the underlying information gap essentially unchanged.
The same is true for claims about community benefits. A corporate blog can explain a company's position, but it isn't a substitute for a record that an independent observer can review. If a company says a facility will create jobs or improve the tax base, those claims should be tied to figures, assumptions, and timelines that can be checked against what happens later.
Ending NDAs is best understood as an opening condition. It creates more space for scrutiny. It doesn't settle the debate, guarantee accurate disclosure, or obligate a community to support a project. Transparency should help people decide - without deciding for them.
Why moratoriums matter even when they are temporary
Some communities and states have introduced temporary pauses on data center approvals or permits while they study impacts. These measures vary in scope. A pause might cover only certain projects, last one to two years, or remain in place until an environmental review is finished.
It's easy to dismiss a short moratorium as symbolic - especially since data centers can take years to go from proposal to operation. But the calendar isn't the only thing that changes. A pause can delay permits, shift assumptions, and give officials time to develop rules for water, power, noise, land use, and public disclosure. Even a temporary delay can affect project economics if funding, zoning, or grid constraints are already tight.
That doesn't mean every moratorium stops construction. Some expire before a project is ready. Others get lifted after a review is complete. The key point is that uncertainty has a cost for developers. If a project depends on predictable approvals, it becomes harder to finance when local rules, public sentiment, or infrastructure availability are still in flux.
A pause can be a governance tool, not a verdict
Moratoriums are sometimes framed as proof that a community has rejected data centers. Sometimes that's true. Other times officials use a pause to study the issue and decide what rules should apply. Those situations aren't the same.
A temporary pause can help a community avoid location-specific decisions without enough information. It can also show that local government hasn't built a process for evaluating fast-growing infrastructure. For a company, that might look like delay. For residents, it can look like a chance to ask questions that weren't part of the original process.
From a business perspective, it's worth taking that difference seriously. If a company treats every pause as irrational obstruction, it may miss the real bottleneck - unclear zoning rules, limited grid capacity, disagreement about water use, or mistrust created by earlier negotiations. Different problems require different responses. A communications campaign won't fix a power constraint. A better engineering pitch won't fix the feeling that the process is illegitimate.
The burden of proof is becoming more local
For years, technology companies could rely on a broad story about innovation and economic growth. That argument lands differently when local costs show up immediately and promised benefits are hard to verify.
The burden is shifting toward project-level evidence:
- What energy demand is expected?
- Is the electricity supply already available, or will new infrastructure be required?
- What are the water assumptions?
- How many permanent jobs are expected?
- What tax revenue or public investment is involved?
- Which commitments are contractual, and which are just projections?
Details will vary from site to site. That's why generic claims about whether data centers are good or bad don't answer the question people are actually facing. The public needs enough information to judge the specific proposal in front of them - not a national-level argument about AI in general.
For a broader view of the infrastructure behind current AI products, see our guide to AI infrastructure and compute. The business case for more capacity depends on demand. The local case for a particular facility depends on its own costs and terms.
A practical framework for earning community trust
Companies can't promise community approval - and they shouldn't treat approval as a communications metric. A legitimate process can end with changes, delays, or rejection. The goal should be to make the decision better informed and the commitments easier to hold accountable.
A workable approach has four stages: disclose early, quantify local effects, create a real feedback loop, and publish follow-through. Each stage fixes a different failure mode. Start too late and suspicion grows. Use vague claims and people can't compare. Hold one-way meetings and concerns don't surface. Lose track of promises after approval and credibility evaporates.
1. Disclose before the proposal feels inevitable
Share the basic facts as soon as you can do it responsibly: site, likely scale, intended use, project stage, expected timeline, and decision points that are still open. If the operator or end user can't be named yet, explain why - and say when that information will be public.
Early disclosure doesn't require pretending everything is final. It does require separating confirmed facts from estimates and openly labeling what's unknown. People can handle uncertainty when it's clear. They're less likely to trust numbers that change later without an explanation.
2. Put local impacts in terms residents can evaluate
Publish project-specific estimates for energy, water, traffic, noise, land use, employment, and any relevant public costs. State the assumptions behind each figure. For instance, if you project water use, explain whether it's an annual estimate, how it varies with operating conditions, and where the water would come from.
If the data isn't ready yet, share a timeline for when it will be. Don't replace missing information with a broad claim that the project will be "efficient" or "sustainable." Those words don't let anyone test the claim.
3. Make engagement two-way and consequential
Public meetings only help if people can ask questions before key decisions are finalized - and if the project team records what it hears and answers it. Offer multiple ways to participate, including written input for residents who can't attend an evening meeting. Publish responses in a format people can find later.
Be clear about what can change. If the site, design, or operating plan is fixed, say that. If there's flexibility - traffic routes, cooling design, construction timing, or community commitments - explain how feedback will be considered. Listening without a real ability to respond turns engagement into a presentation.
4. Track promises after approval
Create a public record of commitments, responsible parties, delivery dates, and progress. If the project promised a specific level of investment or a defined mitigation measure, residents shouldn't have to rely on memory or marketing materials to find out whether it happened.
This is also where companies can distinguish verifiable community benefits from general statements. A promise tied to a permit, an agreement, or a published reporting schedule is more useful than a hope posted on a blog. Local governments should be clear about their own responsibilities, including oversight and enforcement.
A simple test keeps everything grounded: Could a resident who wasn't in the negotiations understand what the company is proposing, what the community is being asked to accept, and how to check whether commitments were kept? If the answer is no, then ending NDAs hasn't gone far enough.
Conclusion
Ending NDAs is a useful change because it challenges the idea that major infrastructure negotiations should happen beyond public view. But transparency isn't a press release, and it isn't a shortcut to approval. It's a process: disclose early, show the assumptions behind local impact claims, make public input matter, and track commitments after the project is approved.
AI companies need infrastructure to build and operate their products. Communities need credible information before they take on the consequences of hosting them. Those needs aren't inherently incompatible - but you can't reconcile them just by saying AI is inevitable or that the benefits will eventually be obvious.
If you're evaluating an AI infrastructure proposal, start with the basics: what's being built, what it will use, who pays, what's promised, and how those promises will be verified. Then judge the project on those terms. That's a better foundation for trust - and a better way to make decisions about the physical systems behind AI.
For more on the decisions shaping responsible AI adoption, explore our AI ethics and accountability resources.