Big Tech is spending billions to win the AI race. Now a growing number of investors are paying for protection in case that spending doesn’t pay off.
As of October 2026, investors are increasingly turning to credit default swaps (CDS) and similar instruments to hedge their exposure to the swelling debt loads of major technology companies. The worry is simple: will future AI revenue arrive fast enough to cover the bills already coming due?
The hedging surge, by the numbers
A credit default swap works like an insurance policy on a borrower. The buyer pays a regular premium. If the borrower defaults, the seller covers the loss.
The price of that premium, called the spread, is a real-time gauge of how nervous the market feels. Wider spreads mean investors see a higher chance that a company won’t pay back what it owes.
October 2026 brought a notable jump in CDS trading volumes, along with record-high spreads on several prominent names, including Oracle, SoftBank, Meta, Nvidia and CoreWeave.
Five-year CDS spreads on Oracle have reached record levels, exceeding 288 to 322 basis points. Meta and Nvidia have both seen spreads climb above 100 basis points.
CoreWeave sits at the sharpest end of the spectrum. Its bonds have traded as low as 77 cents on the dollar, and its CDS pricing implies a greater than 60% probability of default within five years. The company carries tens of billions in liabilities tied to data centers.
A mountain of borrowing, plus the debt nobody sees
Big Tech hyperscalers had issued approximately $466 billion in debt to finance AI infrastructure by mid-October 2026, a sharp rise compared with earlier periods.
Projected AI capital expenditures by the top hyperscalers in 2026 are expected to surpass $690 billion, eating up a large share of their operating cash flow.
Moody’s estimates that $662 billion in lease obligations across five hyperscalers are not yet recorded on balance sheets. Broader analyses suggest total off-balance-sheet liabilities could potentially exceed $1.6 trillion.
Both Moody’s and S&P Global are applying additional scrutiny to these off-balance-sheet liabilities, including future lease commitments, which make the true exposure harder for outsiders to measure.
Why the bond market is getting picky
Rising volatility and a repricing of risk suggest that the easy-money phase of AI financing is meeting some resistance.
The research points to widening spreads and falling bond prices concentrated among companies most directly exposed to heavy AI spending, while core hyperscalers appear to be maintaining stronger financial positions.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.