Nvidia Swaps Top Meta at 128 Basis Points Amid AI Funding Concerns

Key Facts

The cost of protecting against a default on Nvidia debt surpassed that of Meta on October 26, with the five-year credit default swap spread at 128 basis points for the former, versus 121 basis points for the latter. Benzinga reported the figures, citing S&P Global Market Intelligence data, in a reversal from the previous month’s ranking. On September 25, Nvidia’s spread stood at 87 basis points, compared with 98 basis points for Meta. The shift puts the credit risk investors are pricing for Nvidia above Meta’s, after it had been lower about a month earlier. It came amid broader concerns about the cost of financing AI infrastructure and debt-funded spending.

Credit default swaps measure the cost of buying protection against an issuer’s potential failure to meet its obligations. Reuters explained that a swap priced at 100 basis points costs one dollar a year to insure every 100 dollars of debt. When investors become more concerned about creditworthiness, demand for protection rises and spreads widen, potentially increasing a company’s borrowing costs. But a wider spread does not prove a default has occurred; Reuters said companies whose swap spreads have recently risen have not defaulted on their obligations and have not indicated that they might do so. Nvidia and Meta’s figures therefore reflect a market repricing of credit risk, not an announcement that either company has defaulted.

Concerns were sharper for CoreWeave, whose five-year swap spread peaked at 1,167 basis points on October 22, the highest level in the data series covering 12 months up to that point, according to Benzinga, citing S&P Global Market Intelligence data. Benzinga estimated that market pricing implied an approximately 62% probability of default over five years, assuming a recovery rate of 40% that S&P Global uses for the company. This is a probability inferred from prices, not a confirmed forecast of default. Reuters reported on October 15 that CoreWeave spreads had exceeded 1,100 basis points and that its bond maturing in 2031 yielded 16% before falling to about 14%. Reuters also linked the concern to questions about how quickly massive AI investments will translate into profits and whether heavily indebted companies can cover rising interest costs.

Risk pricing also spread to other major technology companies, though at a slower pace than for Oracle and CoreWeave. On October 15, Reuters said swap spreads for Amazon, Alphabet and Meta had reached record levels, while swaps for Nvidia, STMicroelectronics and SpaceX had risen sharply amid concerns about credit pressures tied to AI investments. The concerns center on the financing needed to build data centers and buy additional chips, as borrowing costs rise. In September, AI-related low-rated corporate issuance reached 88 billion dollars through September 30, compared with 20 billion dollars in leveraged finance issuance during the first eleven months of 2025, Reuters reported, citing Goldman Sachs. AI infrastructure financing in the high-yield bond market also reached 40 billion dollars through September 30, versus 12 billion dollars for all of 2025, according to BNP data reported by Reuters.

Sector shares did not follow a similar downward path in the October 26 session despite widening credit spreads. Nvidia shares rose 2.4%, while CoreWeave shares ended the session up 1.8% after sharp trading fluctuations, according to AP. Nvidia closed at 216 dollars and Meta at 630.56 dollars, according to EL7 data. In its fiscal 2027 second-quarter earnings report, Nvidia said revenue was 96.2 billion dollars, up 106% year over year, including 89.0 billion dollars from data centers, up 117%. CEO Jensen Huang said when the results were announced in August that “demand is accelerating” and “AI infrastructure buildouts are happening at full speed.” The figures show the company’s business continuing to grow alongside an increase in the market-priced cost of credit protection.

AI stocks had risen on October 26 after Anthropic announced it had filed for an initial public offering, despite continued investor concern about debt-funded infrastructure spending; Nvidia shares rose about 3%, according to Reuters. A successful offering by a specialized AI company such as Anthropic could reshape the market’s valuation of parts of the industry, including chipmakers and infrastructure companies, and could also divert capital from some major technology companies, according to analysts cited by Reuters. For Nvidia, the company forecast revenue of 108 billion dollars, plus or minus 2%, for the third quarter of fiscal 2027, assuming no data center computing revenue from China. On October 21, the company said its DSX POD modular AI factory design could cut deployment time at a prepared site from six months to 14 days. It added that global deliveries of these units would begin in early 2027.