Key facts
- AI-related debt issuance by low-rated firms has reached $88 billion this year.
- Investors are demanding higher yields from AI borrowers due to unproven revenue streams.
- Data centers are a key source of new money in the high-yield market.
- Borrowing costs for lower-rated AI companies could reach 14% to 15%.
- Collateralized loan obligation (CLO) managers are becoming more cautious on certain AI-linked names.
- Zenith Arc's five-year senior secured notes, issued in August, fell more than seven points below their issue price.
Lenders are increasingly demanding higher compensation to finance AI-related borrowers whose future earnings remain unproven, leading to higher borrowing costs in the riskier segments of the US credit market. This trend comes as higher-rated AI issuers have been actively borrowing and a selloff in Treasury markets pushes yields higher across the board.
AI-related issuance by low-rated firms has reached $88 billion this year, according to Goldman Sachs, with most of it coming from US issuers. In contrast, AI-related issuance in leveraged finance—primarily through junk bonds and loans—totaled $20 billion in the first 11 months of 2025, according to Neuberger Berman data.
Investors are scrutinizing revenue projections, collateral value, and debt capacity of less-established AI borrowers. "High yield people like to know how much cash flow is coming, when that cash flow is coming, and what is the probability that the cash flow doesn't come," said Larry Holzenthaler, senior portfolio manager, fixed income, at Catalyst Funds. He noted that rising investor compensation demands reflect the asymmetric risk where investors are repaid at par if all goes well, but bear losses if it does not.
Despite rising issuance, overall risk appetite for AI-related credit remains subdued, with leveraged finance buyers concentrating on higher-quality borrowers, particularly double-B-rated companies with predictable revenue streams. Data centers, often falling into this higher-quality category, have been crucial for the high-yield market, which is otherwise shrinking. AI infrastructure supply in high yield has reached $40 billion this year, a significant increase from $12 billion for the entirety of 2025, according to BNP data.
However, AI companies are unlikely to push further down the credit spectrum due to the elevated borrowing costs. Issuers near investment grade, such as BB+ rated companies, are already paying yields of around 9% to 10%, with lower-rated borrowers potentially facing costs of 14% to 15%. SoftBank Group, for example, paid yields between 8.625% and 9.75% on its recent notes, which analysts noted are typically associated with significantly lower-rated companies.
Debt issuers are also encountering a more skeptical investor base. Unlike investment-grade investors, buyers of AI-linked high-yield bonds and leveraged loans face constraints like portfolio rules limiting exposure to riskier borrowers. Lotfi Karoui, multi-asset credit strategist at PIMCO, described the proposition for debt investors as fundamentally asymmetric, with largely contractual returns versus significant risks from high debt levels, project delays, and rapid technological changes.
Many AI borrowers require substantial upfront investment before generating reliable cash flow, making their debt harder for collateralized loan obligation (CLO) managers to absorb, especially if leverage increases or ratings deteriorate. CLO managers are reportedly becoming more cautious. The market's skepticism was highlighted by Zenith Arc's August issuance of $2.25 billion in five-year senior secured notes, which priced slightly below par and subsequently dropped significantly below their issue price, according to Pender Fund Management.