Mentioned in This Article
1. AI credit default swaps reach record levels
Credit default swaps covering the debt of several artificial intelligence companies have climbed to record prices, and a J.P. Morgan Private Bank strategist says the shift is changing how lenders judge the sector. In a report dated October 23, Justin Biemann, the bank's global investment strategist, wrote that an AI association by itself no longer reassures creditors. Borrowers must now demonstrate that they can pay for their investment programs, service their debt and eventually turn heavy spending into cash.
The warning arrives as equity markets show signs of strain. The Nasdaq has slipped 6.7% since October 6, falling from 31,360 to 29,263, and AI-linked companies such as Nvidia and Oracle have fallen by more than the index. Taken together, the price action and the credit signals suggest investors are asking a harder question than they were a few months ago: whether the promise of artificial intelligence is arriving on corporate balance sheets.
2. What lenders now expect from AI borrowers
Biemann's central point is that creditors are weighing the business and financial viability of each borrower rather than treating the AI theme as a form of collateral. For much of the recent boom, a company that described itself as part of the AI buildout could often raise money on favorable terms. That assumption is eroding, according to the J.P. Morgan note.
He set out three tests. A company has to show it can fund its investment, meaning it has enough access to capital to keep building. It has to show it can service the debt it has already taken on. And it has to show that the money it is spending will eventually convert into cash flow. In his words, the AI label alone is no longer enough to reassure lenders.
The concern is not evenly spread. Firms with strong existing cash generation face a very different conversation with lenders than those that depend on continuous borrowing to pay for data centers and chips. The second group is where the report focuses its worry.
3. Why Elon Musk's neocloud comment matters
The same theme surfaced publicly on Sunday, October 26, when Elon Musk said that debt financing for AI "neoclouds" is drying up. Neoclouds are companies whose revenue comes almost entirely from operating data centers that rent computing capacity to others. Because such firms typically carry large loans against expensive hardware, they are especially sensitive to any tightening in credit markets.
The remarks from Musk and the analysis from J.P. Morgan Private Bank were produced independently, but they point in the same direction. Both describe a financing environment in which access to borrowed money, rather than demand for AI itself, may become the limiting factor for some operators.
4. What is a credit default swap and how is it priced?
A credit default swap, or CDS, is a form of insurance on debt. The buyer pays a yearly premium to the seller and receives a large payout if the borrower defaults or fails to meet its payments. Investors use these contracts both to protect bond holdings and to take a view on a company's creditworthiness.
The premium is quoted as a spread in basis points, with one basis point equal to one hundredth of a percentage point. A spread of 100 basis points means that insuring $10 million of debt costs $100,000 a year. A wider spread signals that investors are demanding more compensation for the risk that the borrower will not pay. When spreads rise sharply, it usually means the market is pricing in a greater chance of trouble.
That is why the current moves draw attention. The cost of protection on the debt of several AI-linked firms has reached record levels, according to the J.P. Morgan report, which treats the swap market as a real-time gauge of how nervous creditors have become.
5. How much has CDS trading in AI names grown?
Trading volume has expanded dramatically. Data attributed to Morgan Stanley show that weekly CDS activity tied to bonds from Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia averaged more than $9 billion in September. In July, the same weekly figure was below $2 billion, so activity has grown more than fourfold in roughly two months.
A jump of that size can reflect two different motives. Some of it comes from investors who are worried and want protection. Some of it comes from institutions hedging large bond positions in companies that have issued a great deal of debt to fund AI infrastructure. The report argues that, for the biggest names, hedging appears to be the dominant motive rather than a bet on a crisis.
6. The most indebted firms show the sharpest stress
The most troubling readings come from companies that rely on financing markets staying open. Biemann did not name the firm in question, but he described an AI infrastructure company that depends on outside funding and has taken on tens of billions of dollars in debt and other obligations. Its five-year CDS rose to 1,100 basis points, a level at which insuring $10 million of its debt would cost more than $1 million a year.
According to Biemann, pricing at that level implies a probability of default over five years that is near or above 50%. A year earlier, the same measure implied a probability below 30%. The change shows how quickly market perceptions of credit risk can move when funding conditions tighten and spending continues at a high rate.
Some established infrastructure borrowers have also seen their protection costs climb, reaching toward 300 basis points. By contrast, protection costs across the broad investment-grade market remain comparatively contained, which indicates the stress is concentrated rather than systemic.
7. Warning lights, not countdown clocks
Biemann was careful to separate a warning from a forecast. He pointed to Lehman Brothers as the historical example of what a CDS signal can precede. The investment bank's CDS averaged below 100 basis points in the third quarter of 2007 and exceeded 1,400 by mid-September 2008, when the firm filed for bankruptcy and unsettled global markets. Biemann presents the episode as evidence that swaps can serve as an early warning.
At the same time, he urged investors to read today's readings as "warning lights rather than countdown clocks." The strongest AI leaders, those with substantial cash flow, pay roughly 50 to 120 basis points for five-year protection. Those prices, he said, still look more like the price of portfolio insurance than a forecast of distress.
The distinction matters for how the data should be used. Elevated spreads on a few leveraged operators tell investors where the vulnerabilities sit. They do not, on their own, indicate that the sector as a whole is heading for a collapse comparable to 2008.
8. Why debt matters so much to the AI buildout
Building artificial intelligence capacity is capital intensive. Data centers, specialized chips and the power infrastructure to run them require large sums spent well before the revenue they enable appears. Companies with big profits can pay for much of this from their own cash. Others borrow, issuing bonds or taking loans secured against equipment, and then rely on future customer payments to repay the lenders.
That arrangement works as long as two conditions hold. Lenders must remain willing to refinance or extend new credit, and revenue must grow quickly enough to cover interest and principal. If either condition weakens, a borrower that is still spending heavily can find itself squeezed from both sides. Rising default insurance prices are one of the earliest places where such a squeeze becomes visible, because swap traders reprice risk faster than ratings agencies or annual reports can.
For lenders, the practical consequence is closer scrutiny. Instead of accepting a growth narrative, they ask for evidence: contracts with creditworthy customers, a clear schedule for how borrowed money will be repaid, and a path from capital spending to operating cash. Borrowers who cannot answer those questions may face higher interest costs, tighter terms or reduced access to funding, which in turn limits how quickly they can expand.
This is also why the report separates the largest, cash-rich technology companies from the more financing-dependent operators. The first group can absorb a period of expensive credit because internal cash flow supports their spending. The second group has far less room, and a prolonged period of costly or unavailable borrowing would test its business model directly.
Investors tracking the sector therefore have two separate questions to answer. The first is whether demand for AI services keeps growing. The second, which the swap market is now pricing, is whether each borrower can pay for that growth without depending on credit conditions that may not last.
9. Why this matters for Bitcoin and crypto markets
The link to digital assets runs through risk appetite. Artificial intelligence has been one of the main forces encouraging investors to take risk across markets, crypto included. Bitcoin (BTC) and Nvidia both reached their lows within weeks of ChatGPT's debut in late 2022, and both have rallied substantially since then. That shared history means a deeper downturn in AI-related assets could spill into crypto through reduced appetite for speculative holdings.
The transmission is not automatic, and the report does not claim that credit stress in AI will produce losses in crypto. It does suggest that traders watching Bitcoin should keep an eye on the credit market for AI borrowers, because a sustained rise in default insurance costs would tighten the financial conditions that support risk assets generally.
What to watch next is whether spreads on the most leveraged AI operators keep widening, and whether those companies can show cash generation that justifies their borrowing. Evidence of either would shape how lenders treat the sector in the months ahead.