Corporate bond investors have grown more selective toward debt from AI companies and cloud hyperscalers, as a surge of new issuance has pushed those credit spreads wider even as the broader investment-grade market has not fully followed.
J.P. Morgan Private Bank, using data as of early September, said U.S. dollar public AI debt sales had already reached $260 billion year to date. “We’re seeing signs investors are worried about the rapid increase in debt issuance by AI hyperscalers and AI-related companies,” the bank wrote. It said the widening had not shown up as broad spread widening across the overall investment-grade index, suggesting the impact had been contained to hyperscalers.
Secondary market reports on Sept. 22 described highly rated corporate credit as splitting, with caution toward AI-related issues and continued demand for traditional financial and industrial borrowers. Those accounts said managers were focused less on the creditworthiness of major technology companies and more on the scale of supply. The original wire report could not be independently retrieved for this story.
That supply has risen sharply. IESE Business School said in a June 30 analysis that gross bond issuance by the five largest hyperscalers climbed from about $20 billion in 2024 to $109 billion in 2025, including Meta’s $30 billion sale on Oct. 30, 2025, and Alphabet’s $25 billion sale in November 2025. With data as of mid-May, IESE put year-to-date hyperscaler issuance at about $152 billion and said some $300 billion was expected in 2026.
"Spreads have so far absorbed the supply remarkably well, but not without strain," IESE wrote. "Even top-tier tech giants like Alphabet and Meta are paying a premium to entice buyers due to the sheer volume of AI debt." The school also said AI capital spending was outrunning operating cash flow for several issuers.
Investors disagree on whether wider spreads are mainly a supply jam or a change in how the market values the borrowers. J.P. Morgan Private Bank said it saw "selective opportunity in investment grade hyperscaler debt as recent spread widening looks more technical than fundamental." Apollo chief economist Torsten Slok has argued the move is about credit fundamentals. Economist Brad DeLong, citing Slok on Sept. 16, wrote that the widening in hyperscaler CDS "is not driven by dealer hedging of new issuance" and that markets are repricing "a debt-financed AI capex cycle with rising leverage, negative free cash flow and uncertain payback on depreciating assets," while noting spreads were still only about 0.6 percentage point.
If investment-grade credit is splitting, the AI buildout is no longer just an equity story. It changes who can borrow cheaply and how easily the next wave of data-center and model financing clears — including debt-heavy funded bets that need credit markets to keep absorbing supply. The open question is whether wider AI-related spreads are a temporary traffic jam or an early price on weaker cash-flow payback and ROIC opacity that equity markets have been slower to force.