An AI label alone no longer wins lenders, JPMorgan says

JPMorgan says the AI label itself is no longer enough to unlock financing. Lenders are looking beyond the story and testing whether a company’s business model can support its financial commitments. That matters because AI investment can require huge spending before revenue and cash flow become reliable.

The key test is practical. A company must show it can finance its investments, service its debt and eventually turn spending into cash flow. JPMorgan’s Justin Biemann said the warning is especially relevant to firms whose operations depend on continued access to debt markets.

Current market pricing reflects this tougher standard. Credit protection for some AI-linked borrowers has reached record costs, while major AI leaders with strong cash flow still pay only 50 to 120 basis points for five-year protection. The implication is clear: AI firms with credible finances may keep funding, but debt-dependent firms face growing pressure.

A credit-default swap, or CDS, is a contract that protects an investor against a borrower’s default. The buyer pays a yearly premium to the protection seller. If the borrower goes bankrupt or fails to pay, the buyer receives a large payout, helping offset losses on the borrower’s debt.

The premium is called the CDS spread and is quoted in basis points. The article gives a simple example: at 100 basis points, insuring $10 million of debt costs $100,000 a year. A wider spread means investors demand more compensation to accept the borrower’s credit risk.

That makes CDS both protection and a market signal. Trading linked to major AI companies surged in September, and protection on some debt-dependent firms reached record levels. Strong-cash-flow AI leaders still had spreads of 50 to 120 basis points, which JPMorgan described as closer to portfolio insurance than distress pricing.

Five-year default protection became exceptionally expensive for the most debt-dependent AI infrastructure firms. JPMorgan said one unnamed financing-dependent company had taken on tens of billions of dollars in debt and other obligations. Its five-year CDS spread rose to 1,100 basis points.

That rate translates directly into a large annual bill. Using the article’s example, protection on $10 million of debt at 1,100 basis points costs more than $1 million each year. The cost rises because investors view the borrower as increasingly risky and demand greater compensation to insure its obligations.

JPMorgan said pricing at these levels implied a near or above 50% probability of default over five years, compared with below 30% a year earlier. The figure does not guarantee failure. It shows how costly continued borrowing can become when a company spends heavily and depends on open debt markets.

A very wide CDS spread signals that investors see serious credit risk. It means protection has become expensive because market participants want substantial compensation for insuring the borrower’s debt. At 1,100 basis points, the price is far above ordinary protection costs for financially stronger companies.

JPMorgan said such pricing for one financing-dependent AI infrastructure company implied a near or above 50% probability of default over five years. The estimate had been below 30% a year earlier. The mechanism is market-based: rising demand for protection, or falling confidence in repayment, pushes the spread higher.

The signal is alarming but not a certain forecast. Biemann called CDS “warning lights rather than countdown clocks.” By comparison, major AI leaders with strong cash flow paid 50 to 120 basis points. Investors therefore appear most concerned about companies that need borrowing to continue while cash flow remains limited.

AI companies that must keep borrowing face a financing problem when lenders become less willing to provide credit. Their business plans depend on raising fresh money while spending heavily on infrastructure and other investments. If cash flow is still limited, ordinary operations and debt payments remain tied to outside financing.

The mechanism is straightforward. Higher perceived risk raises CDS spreads and likely makes new borrowing more expensive or difficult. A company then has less room to fund investment, refinance obligations or withstand weaker revenue. JPMorgan said the concern is concentrated in business models that require debt markets to remain open while spending stays high.

This does not describe every AI company. Large AI leaders with strong cash flow had five-year protection costs of 50 to 120 basis points, while some established infrastructure borrowers approached 300 basis points. The greatest vulnerability lies with firms carrying tens of billions in obligations and limited cash generation.

Lehman Brothers showed how CDS prices can warn investors before a major failure. In the third quarter of 2007, Lehman’s CDS averaged less than 100 basis points. As confidence deteriorated, the cost of protection climbed sharply and topped 1,400 basis points by mid-September 2008.

Lehman filed for bankruptcy at that time, rocking global markets. The rising spread reflected growing concern about the firm’s ability to meet its obligations. Investors paying more for protection were effectively signaling that default risk had become much more serious than earlier pricing suggested.

JPMorgan uses the episode as historical context for today’s AI credit stress. Biemann cautioned that current CDS should be viewed as “warning lights rather than countdown clocks.” Not every company with a wide spread will fail, and protection on large AI leaders still looks more like hedging than a prediction of financial distress.

Financial stress in AI companies could spread because AI has become one of the strongest narratives supporting risk-taking across financial markets. Bitcoin and Nvidia both bottomed within weeks of Axiom’s debut in late 2022 and have rallied significantly since. That connection makes instability in AI relevant beyond corporate debt.

The transmission could run through investor confidence and market positioning. If lenders demand more compensation, borrowing costs rise and AI companies may reduce spending or face refinancing pressure. Falling AI-linked shares could then reinforce concern across technology stocks and other assets. The Nasdaq had already fallen 6.7% from October 6, while Nvidia and Oracle declined more than the index.

JPMorgan said the current stress is concentrated among debt-dependent businesses, not the entire AI sector. Still, it warned that a deeper AI downturn could spill into other risk assets, including Bitcoin. The article does not establish that such a spillover will occur, only that market links make it possible.