Amazon and Microsoft Pause AI Data Center Plans Despite U.S. Push

Amazon and Microsoft are both pausing or slowing key AI data center projects—even as the U.S. government calls for an aggressive buildout of domestic AI infrastructure.

Amazon Web Services (AWS) has paused some large-scale data center leasing talks—especially abroad—amid rising economic uncertainty and growing scrutiny over how billions earmarked for AI infrastructure, including costly Nvidia chips, are being spent. Analysts at Wells Fargo described the move as part of a broader trend of hyperscalers—major cloud providers like Amazon, Microsoft, and Google—becoming more cautious with power-intensive leasing commitments through 2026.

While Amazon stressed this is “routine capacity management,” the pause mirrors Microsoft’s more public retreat from several data center initiatives in the U.S. and Europe. Microsoft recently scrapped projects representing 2 gigawatts of power, citing a mismatch between infrastructure growth and actual demand. Some of these shelved projects, such as one in Ohio's Licking County and another in Wisconsin, will now be repurposed or held for future use.

“In recent years, demand for our cloud and AI services grew more than we could have ever anticipated, and to meet this opportunity, we began executing the largest and most ambitious infrastructure scaling project in our history” said Noelle Walsh, Microsoft’s head of cloud operations. “By nature, any significant new endeavor at this size and scale requires agility and refinement as we learn and grow with our customers. What this means is that we are slowing or pausing some early-stage projects.”

"This is routine capacity management, and there haven't been any recent fundamental changes in our expansion plans," said Kevin Miller, vice president of Amazon Web Services Global Data Centers.

Despite these slowdowns, both companies continue to pour billions into AI development. Amazon CEO Andy Jassy recently defended the company's heavy investment in generative AI to stay competitive, while Microsoft confirmed it still plans to spend over $80 billion on infrastructure this year.

A Federal Push Meets Corporate Caution

These decisions arrive just as the Trump administration intensifies efforts to bolster America’s AI infrastructure. The Department of Energy has identified 16 new sites for future data centers on federal land, framing the initiative as essential to national competitiveness in the “global race for AI dominance" with Chris Wright, Secretary of Energy, calling the race “the next Manhattan Project,” targeting initial operations by the end of 2027. These centers are expected to support machine learning, AI systems, and massive cloud storage.

What This Means

The decision by Amazon and Microsoft to pause or slow AI data center projects marks a surprising recalibration at a moment of intense global competition. As the U.S. government accelerates efforts to solidify AI leadership—mapping out federal sites for new data infrastructure—the tech giants driving innovation are becoming more cautious with how and where they scale.

This shift reflects growing investor concerns about the return on massive AI spending, especially as emerging competitors like China's DeepSeek demonstrate comparable technology at a fraction of the cost. Meanwhile, the race to secure power, land, and chips for next-gen AI models is becoming more complex and expensive, forcing companies to prioritize flexibility over speed.

Ultimately, the pause isn't a retreat from AI—it’s a strategic regrouping. But the disconnect between national ambition and corporate execution could create friction as the U.S. seeks to lead the AI frontier.

In the race for AI dominance, it’s not just about who builds fastest—it’s about who builds smart, stable, and ready for what comes next.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.