Tech giants are transforming swaths of the American heartland into energy-hungry AI campuses, financing the push with an unprecedented debt binge that is rippling through credit markets. OpenAI, at the center of the buildout, is advancing an $850 billion multiyear “Stargate” data-center program with partners from Nvidia and Oracle to AWS, even as power availability and grid connections emerge as the hard constraint. Meta, Google, Microsoft, Amazon, and Musk’s xAI are racing to plant hyperscale sites from Texas and Louisiana to Wisconsin and Indiana—projects that rival mid-sized cities in electricity demand.
Capital spending is ballooning: hyperscalers are on pace to invest about $443 billion this year, rising to an estimated $602 billion in 2026, with roughly three-quarters directed to AI infrastructure. To bridge the gap, companies raised $121 billion in new debt this year alone, while CDS spreads widened and analysts warned of echoes of the dot-com fiber glut. OpenAI’s web of commitments—spanning chips, cloud, and custom silicon—has prompted concerns about a “circular” AI economy and the risk of stress transmission if demand falters. Yet executives insist that inference-driven usage is compounding, power remains the bottleneck, and the payoff justifies the leverage. The next phase hinges on energized real estate, permitting, and whether today’s scale thesis survives the credit cycle.
The year AI titans, powered by massive borrowing, began reshaping America
- Trending
- Comments
- Latest
