Overview
- As of July 10, 2026, companies tied to AI infrastructure hold roughly $350 billion of related debt while the largest spenders plan about $725 billion of capital expenditure for 2026, leaving cash flow under strain.
- Firms are turning to very long‑dated bonds, delayed‑draw loans and multi‑currency deals to match decades‑long data‑center costs with long‑term financing.
- Recent large raises showed softer demand, with an Amazon bond sale this month flagged as unusually weak and traders selling existing tech bonds to make room for new supply, which has pushed secondary spreads wider.
- The U.S. Senate has opened an investigation into opaque financing structures and disclosure to lenders, and ratings agencies have warned higher leverage could pressure some firms’ credit profiles.
- Consultants and banks project cumulative AI data‑center spending could reach into the trillions by 2030, a shift that could crowd out European borrowers, change where corporate debt is issued, and force firms to cut buybacks or other shareholder returns.