Amazon Warns Data Center Moratoriums Risk Losing the AI Race

Amazon Web Services CEO Matt Garman published a sweeping essay today arguing the United States "can't afford to lose" the race for AI dominance, warning that the more than 100 local data center moratoriums under consideration across the country could hand the advantage to rivals. The roughly 3,000-word piece frames the AI infrastructure buildout as the largest since the interstate highway system and announces a new "Amazon Data Center Commitment" alongside a $1 billion community program.

What Amazon announced

The essay, published on Amazon's corporate news site, codifies the company's data center practices into a four-part commitment: protecting local electricity rates, investing in jobs and economic benefit, ending nondisclosure agreements with government agencies, and building more efficiently. Amazon says its data centers run at a power usage effectiveness of 1.14, below the 1.25 industry average, and that it aims to be "water positive" by 2030.

Alongside the commitment, Garman unveiled "Built Together," a community investment framework that will add more than $1 billion over five years to education, job training, and energy affordability programs in data center communities.

Why Amazon is speaking now

The push comes as public support for AI data centers has fallen sharply in polling, with local backlash mounting over power demand, water use, and environmental impact. Garman directly rebutted what he called "myths," arguing that data centers account for about 0.5% of US industrial water use and that backup diesel generators sit idle 99.9% of the time.

He went further, claiming there are "widespread reports of various countries intentionally seeding misinformation in the U.S. about data centers to trick us into slowing down." If the moratoriums pass, he warned, "the U.S. could be writing its own losing ticket to this race."

Why it matters for builders

For AI builders and automation teams, the story is ultimately about compute. Every frontier model, agent platform, and workflow depends on the data center capacity Amazon and its rivals are racing to build. Where that infrastructure goes, and whether local resistance slows it, will shape the cost and availability of AI services for years. The announcement also signals a shift: after months of safety scrutiny, cloud providers are now aggressively defending the physical buildout that makes AI possible.