AI Debt Crisis: Bond Yields Spike as Infrastructure Costs Outpace Revenue

Key points

• CoreWeave's five-year credit default swap spread reached a record 1,195 basis points, meaning it costs $119,500 annually to insure $1 million of its debt. Its October 2031 bond yield hit 16%, and some bonds traded at 80% of face value, entering 'distressed' territory.
• Oracle's five-year CDS spread neared 300 basis points, a company record, following a credit rating downgrade from S&P Global that placed it just one notch above 'junk' grade. Its November 2032 notes yielded 7.24%.
• Gartner forecasts that AI providers will spend $7 trillion on infrastructure through 2030 but generate no more than $3.6 trillion in related revenue, with over $4 trillion at risk of write-downs by 2032. More than 80% of specialized AI cloud providers ('neoclouds') are predicted to cease independent operations by 2030.
• Global AI-related debt issuance plunged to $23 billion in September 2026, down from $113 billion in June. Banks like JPMorgan are tightening credit to avoid concentration risk, and approximately $18 billion in bank loans for an Oracle data center campus recently entered stressed territory.
• Nvidia CEO Jensen Huang publicly reassured investors that CoreWeave's 'financial strength is improving every day' amid the sell-off, though bond markets remained skeptical. The 10-year U.S. Treasury yield is near 5%, increasing the cost of borrowing for AI firms.

Background

This crisis follows a recent shift in U.S. Treasury rhetoric, where Secretary Scott Bessent conceded that bond markets defied his previous aggressive stance, leading to 10-year yields reaching their highest level since 2007. Earlier this month, CoreWeave shares had already plunged 15% as bond yields spiked, with the CEO dismissing the market reaction as 'irrational.' The current escalation marks a broader move from cash-funded to debt-funded AI expansion, with a third of capex now coming from debt, raising concerns about the sustainability of the AI boom.

How outlets are covering it

The Motley Fool emphasizes the immediate financial distress, highlighting record-high credit default swap spreads for CoreWeave and Oracle, and the macroeconomic pressure from rising oil prices and inflation. Axios focuses on the structural mismatch between spending and revenue, citing Gartner's forecast of a $3.6 trillion revenue gap against $7 trillion in spending, and notes that banks are actively pulling back due to concentration risks. The Financial Times, while behind a paywall, frames the situation as a 'Darwin moment' for AI borrowers, suggesting that the market is now testing the viability of these debt-fueled projects. All sources agree that the gap between infrastructure costs and returns is widening, but they differ in emphasis: Motley Fool highlights specific bond metrics, Axios focuses on long-term economic viability, and the FT suggests a broader market correction is underway.

Why it matters

The AI boom's reliance on debt to fund infrastructure creates systemic risks for the broader economy and stock market. If AI projects fail to generate sufficient revenue to cover interest and operating costs, it could trigger a wider financial crisis. The shift from cash to debt funding, combined with rising borrowing costs and bank pullbacks, threatens the sustainability of the AI industry and could lead to significant write-downs and corporate failures, impacting markets and the broader economy.

What to watch

Investors will watch for further spikes in credit default swap spreads and bond yields for AI firms, particularly CoreWeave and Oracle. Banks may continue to tighten credit, making it harder for AI companies to secure financing. The market will assess whether AI providers can generate sufficient revenue to justify their debt loads, or if a wave of write-downs and corporate failures will follow. The 10-year U.S. Treasury yield and inflation data will also influence borrowing costs for AI firms.