Amazon drops NDAs for data center deals amid growing public backlash

Amazon Web Services has officially stopped using nondisclosure agreements (NDAs) when negotiating with government agencies for new data center projects. AWS CEO Matt Garman made the announcement in a blog post published in early October 2026, aiming to reduce community suspicion and counteract a rising tide of local regulatory pushback against infrastructure expansion.

What happened

The decision comes as tech companies face intensifying scrutiny over the environmental and social impact of their physical infrastructure. Garman’s statement was part of a broader effort to argue that data centers benefit communities, directly addressing the lack of transparency that has fueled public anger. Environmental activist Erin Brockovich recently identified transparency as the primary complaint from residents, noting a pattern where permits are secured before neighbors are informed, often because local officials signed NDAs.

This opacity has triggered concrete legislative responses. New York implemented a one-year moratorium on permits for large data centers, and Garman noted that more than 100 similar moratoriums are currently under consideration across the United States. He warned that if these measures become law, the U.S. risks falling behind in the global technology race, with consequences that could last for generations.

Beyond the NDA policy change, Garman attempted to debunk four common myths regarding data center operations: excessive water consumption, rising electricity costs, high pollution levels, and a lack of community benefit. He cited internal Amazon reports stating that direct data center water use accounts for only 0.5% of industrial water usage in the U.S., far less than agriculture or recreational industries. However, independent scientists argue that without federal reporting requirements, these claims remain difficult to verify independently.

How it works

The core of the controversy lies in how data center impacts are measured and reported. Tech companies typically report "direct" operational metrics, such as water used for cooling towers on-site. Nvidia, for instance, claims its new cooling systems eliminate nearly all internal water usage. Critics and scientists point out that this view ignores the indirect water footprint required for electricity generation and semiconductor manufacturing, which are integral to the data center supply chain.

Regarding energy costs, Garman argued that rate increases are not caused by data centers themselves but by aging grids that failed to expand ahead of demand. He stated that in many states with significant data center presence, electricity rates have grown more slowly or even decreased. Conversely, an independent watchdog recently attributed a 76% year-over-year price increase on America’s largest electrical grid primarily to data center load.

On pollution, Garman addressed concerns about permitted emission caps, such as a Texas facility allowed to release 33 million tons of carbon dioxide annually. He clarified that backup generators, which hold these permits, remain idle 99.9% of the time, running only about 10 hours per year for maintenance testing. Despite these explanations, trust remains low, with Anthropic CEO Dario Amodei describing the current sentiment as a fundamental crisis of trust where the public assumes malicious intent from tech firms and governments.

Key details

• AWS has ceased using nondisclosure agreements with government agencies for data center project approvals.
• More than 100 data center moratoriums are currently being considered across the United States.
• New York has enacted a one-year moratorium on permits for large data centers.
• Amazon claims direct data center water consumption is just 0.5% of total U.S. industrial water usage.
• Backup generators at data centers are reportedly idle 99.9% of the time, running roughly 10 hours annually.
• Amazon states it has contributed over $1 billion to U.S. communities with data center presence in the last three years.

Why it matters

For software engineers and technical leaders, the stability of cloud infrastructure is paramount. The proliferation of local moratoriums threatens the pace at which new compute capacity can come online, potentially leading to resource constraints or higher costs for AI and machine learning workloads. If the U.S. restricts data center growth while other nations expand, it could alter the global competitive landscape for AI development, affecting everything from model training times to service availability.

Furthermore, the shift away from NDAs signals a new era of operational transparency that developers may need to navigate. As communities demand more accountability, tech companies may face stricter local regulations regarding energy sourcing and water usage. Understanding these geopolitical and regulatory shifts is crucial for long-term architectural planning, especially for organizations building energy-intensive AI products that rely on consistent, scalable power and cooling resources.

What you can do

• Monitor local regulatory developments in regions where your primary cloud providers operate to anticipate potential service changes.
• Evaluate the energy efficiency of your own applications to reduce dependency on rapidly expanding infrastructure.
• Stay informed about the sustainability reports of your cloud providers, looking for third-party verified data rather than just marketing claims.
• Consider multi-region deployment strategies to mitigate risks associated with local moratoriums or grid instability.
• Engage with internal sustainability teams to align product roadmaps with emerging environmental compliance standards.
• Diversify cloud vendor relationships to avoid over-reliance on a single provider’s infrastructure rollout schedule.