AI data center boom remakes U.S. landscape as tech giants borrow billions

Across the American heartland, technology companies are converting cheap land and shuttered industrial sites into vast AI data-center campuses, reshaping local economies and stressing power grids. OpenAI’s Stargate complex in Abilene, Texas, is a visible example: one data center is online, a second is nearly complete, and the campus could scale past a gigawatt of capacity.

OpenAI says individual Stargate sites cost roughly $50 billion and could total about $850 billion in spending across the company’s planned buildout. Other hyperscalers are racing to match that scale. Meta is building a four-million-square-foot Hyperion site; Google and Amazon have multibillion-dollar campuses under construction; Microsoft is investing more than $7 billion in a major data center; and Elon Musk’s supercomputer projects in Memphis are expanding into multiple buildings with their own power purchases.

Analysts and companies say the wave of construction reflects an expectation that more compute will keep producing more capable AI systems. Investors and executives call it a transformation on the scale of electrification or the early internet, while some warn it could be a debt-fueled bubble if demand softens.

The capital spending is already enormous. The top five hyperscalers are on track to spend roughly $443 billion this year, with forecasts of $602 billion next year, and most of that expected to go into AI infrastructure.

Key Topics

AI, Tech, Data Centers, Ai Infrastructure, Debt, Openai, Hyperscalers