AI Debt: Investors Demand Premiums, Favor Traditional Issuers

The U.S. investment-grade bond market is dividing between AI-linked borrowers and more traditional issuers, with investors demanding greater compensation from hyperscalers while strongly backing financial, industrial, pharmaceutical and insurance companies. The shift reflects concerns about the scale and unpredictability of borrowing to fund data centers, chips and other AI infrastructure—not expectations that major technology companies will default. Goldman Sachs projects hyperscaler debt issuance will reach about $420 billion next year, a 60% increase from its 2026 estimate, while overall U.S. corporate issuance through August reached $1.9 trillion, up 30% year over year. AI-related issuers have faced wider spreads and larger pricing concessions, as investors manage concentration risk and question the returns on massive capital expenditures; some deals from highly rated companies are pricing more like lower-rated debt. Investors still have cash to deploy but are increasingly seeking opportunities outside the largest technology companies.