As the AI infrastructure buildout pushes into the riskiest corners of U.S. credit markets, investors are demanding higher yields and stricter proof of cash flow to finance low-rated borrowers with unproven revenue models.
Key Financial Metrics & Debt Trends:
Explosive Issuance Growth: Low-rated AI issuance surged to $88 billion in 2026, up from just $20 billion in all of 2025 across U.S. leveraged finance (Goldman Sachs & Neuberger Berman data).
Data Center Expansion: AI infrastructure in high-yield credit reached $40 billion in 2026 (vs. $12 billion in 2025), keeping overall junk bond issuance flat (BNP Paribas data).
Surging Borrowing Costs: Near-investment-grade (BB+ rated) AI issuers face 9% to 10% yields, while lower-rated borrowers face punishing borrowing costs of 14% to 15%.
SoftBank Benchmark: SoftBank (BB+ rated) recently paid high yields of 8.625% to 9.75% on multi-tranche debt issues to fund its AI commitments.
Secondary Market Stress: Zenith Arc’s $2.25 billion data center bond issue dropped over 7 points below issue price shortly after launch, illustrating growing investor caution.
Market Dynamics & Asymmetric Risk: Credit investors face asymmetric risk—capped upside from contractual coupons against full downside loss if technology or project cash flows fail. As a result, CLO managers and high-yield buyers are clustering strictly around double-B issuers with tangible assets and long-term lease backings.
A clear sign that debt markets are demanding concrete cash flows and fiscal discipline as the AI buildout expands beyond mega-cap balance sheets!
