AWS Drops NDAs Amid Growing Data‑Center Opposition

Community groups and officials across the United States have raised alarms over the speed and secrecy with which large data‑center campuses are approved. More than a hundred local moratoriums are under consideration, and New York has already imposed a one‑year permit freeze.

In a blog post, AWS CEO Matt Garman argued that many criticisms rest on four myths: excessive water use, higher electricity rates, massive pollution, and a lack of community benefit. He cited an internal report showing direct data‑center water consumption at just 0.5 % of U.S. industrial use and claimed backup generators run only about ten hours per year.

Independent analysts, however, point out that those figures omit water and energy consumed in power generation and chip fabrication. A watchdog recently linked data‑center demand to a 76 % year‑over‑year price jump on the nation’s largest grid, and the NAACP is suing over permitted emissions that could exceed any single power plant.

Why it matters for GPU / AI infrastructure

GPU‑heavy workloads depend on predictable power, cooling, and fast permitting. When opacity slows approvals or triggers moratoriums, cloud providers like AiGpu face delays and cost uncertainty. Transparent engagement with regulators and communities can accelerate site readiness and secure the reliable energy supplies that AI training clusters require.

By abandoning NDAs, AWS signals a shift toward openness that could set a new industry standard — one that benefits every operator building the next generation of AI compute capacity.