Investors are increasingly questioning the financial sustainability of the artificial intelligence boom as the cost of insuring against defaults by technology companies reaches record highs, exposing growing anxiety over the debt being accumulated to build AI infrastructure.
Credit default swap (CDS) spreads on several technology companies rose sharply on Thursday as hotter-than-expected US inflation data and expectations of higher interest rates intensified concerns about whether the enormous capital expenditure on artificial intelligence will generate sufficient returns to justify the borrowing behind it.
The pressure was most pronounced at CoreWeave, the heavily indebted AI cloud-computing provider whose five-year CDS spreads surged above 1,100 basis points, according to S&P Global Market Intelligence data. The yield on its 2031 bonds climbed to as high as 16% before retreating to around 14%, signaling the premium investors are demanding to hold its debt amid mounting concerns about its financial position.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
The developments suggest that the AI investment debate is moving beyond questions about valuations and future profitability to the ability of companies to service debt while financing the construction of data centers and computing infrastructure.
CoreWeave, a loss-making specialist cloud provider with nearly $50 billion in debt, according to LSEG data, has become a focal point for these concerns. Its financial exposure illustrates the risks facing companies that must commit enormous sums to infrastructure before they can establish whether the resulting revenue will generate adequate returns after interest expenses and other operating costs.
“CoreWeave is quickly becoming the AI trade’s canary in the coal mine,” said Kathleen Brooks, research director at XTB.
“Its bonds have sold off sharply … this is a sign that investors might be going off the AI investment narrative, which has included massive debt issuance to fund AI ambitions.”
The deterioration in credit markets came even as AI developer Anthropic reported a substantial increase in revenue, demonstrating that strong commercial growth among leading AI companies has not been enough to eliminate concerns about the financing structures supporting the wider industry.
Credit Markets Expose The Cost of The AI Infrastructure Boom
Credit default swaps protect against the risk that a borrower will default on its debt. Their spreads represent the annual cost of that protection, expressed in basis points, with wider spreads generally indicating that investors perceive greater credit risk.
CoreWeave’s five-year CDS spread moving above 1,100 basis points is therefore a significant warning signal. The source data indicate that the implied five-year default probability has risen to approximately 64%, compared with around 30% a year earlier.
Such estimates are model-dependent and are not guarantees that a company will default. Nevertheless, the increase shows how dramatically the market’s assessment of CoreWeave’s credit risk has changed.
The bond market is sending a similar message. A yield of around 14% on CoreWeave’s 2031 debt means investors are demanding exceptionally high compensation for lending to the company. The rise in yields also increases the cost of issuing new debt and could complicate refinancing if existing obligations mature before the company generates sufficient cash flow.
The pressure is not limited to smaller AI infrastructure providers. Oracle, whose borrowing to support its expanding cloud and AI ambitions has drawn investor scrutiny, saw its CDS spreads approach 300 basis points. The increase this week reached approximately 31 basis points.
CDS spreads on Nvidia, chipmaker STMicroelectronics and SpaceX, which also offers AI services, rose sharply. Amazon, Alphabet and Meta also recorded their highest CDS spreads, according to LSEG and S&P Global Market Intelligence data, although their increases were considerably slower than those of Oracle and CoreWeave.
The AI ecosystem now includes businesses with substantially different financial profiles. Large technology platforms have established operations and revenue streams that can help finance investment, while specialist infrastructure providers can face greater pressure when expansion depends heavily on borrowing and continued access to capital markets.
The latest moves do not establish that the entire AI sector is approaching a credit crisis. They do, however, indicate that investors are differentiating more sharply between companies based on their debt exposure and capacity to absorb higher financing costs.
Rising Interest Rates Threaten Highly Leveraged AI Businesses
Hotter-than-expected US inflation has added another complication by strengthening concerns that interest rates could remain elevated or rise further. Higher rates can affect AI infrastructure companies through several channels, including more expensive refinancing, greater interest expenses, and a higher return threshold for investors funding long-term projects.
Data centers require substantial upfront investment in computing equipment, power infrastructure and facilities. Companies must commit capital before receiving the full revenue associated with the resulting capacity, creating a timing mismatch between expenditure and returns.
That mismatch becomes more difficult to manage when projects are financed with debt. Companies must meet interest and principal obligations regardless of whether demand develops as quickly as expected or whether customers generate sufficient revenue to justify the investment.
For loss-making businesses such as CoreWeave, the issue is particularly acute. Rapid revenue growth may not be enough if operating costs, capital expenditure and debt-servicing requirements continue to rise faster than cash generation.
Oracle faces a different but related challenge. Its scale and established business operations distinguish it from specialist providers, but its substantial borrowing has made investors more sensitive to the pace and profitability of its AI expansion.
The broader concern is that the financing model underpinning the AI boom could become more vulnerable as investment shifts from being supported primarily by companies’ existing cash flows to relying more heavily on debt issuance.
“This is not the first time we’ve had a sizeable market wobble on the back of AI fears, but as we move to a capex cycle funded out of debt rather than simply free cash flow, we have to increasingly respect the credit price action,” said Deutsche Bank strategist Jim Reid.
“AI financing costs aren’t dramatic overall, but there are worrying themes at the edges,” he added.
Reid’s assessment captures the distinction between the overall technology sector and its more leveraged participants. Even if the largest AI companies can sustain heavy investment, suppliers and infrastructure operators carrying substantial debt may face tighter financial conditions if expected returns take longer to materialize.
Anthropic’s Growth Fails to Settle Investor Concerns
Anthropic’s reported revenue growth offered a reminder that demand for AI services remains strong, but it did little to calm concerns about the sector’s financing risks.
According to an investor who attended a pre-IPO meeting with the company’s leadership in San Francisco on Wednesday, Anthropic’s third-quarter revenue rose 45% from the previous quarter to $16.8 billion. The company is targeting a valuation of $2 trillion.
The revenue increase provides evidence of the commercial momentum enjoyed by a leading AI developer. However, the performance of an AI model developer does not necessarily resolve questions about the profitability of the infrastructure businesses supplying its computing capacity.
Revenue growth across the AI ecosystem does not automatically translate into attractive returns for every company investing in it. Infrastructure providers must account for equipment costs, power consumption, financing expenses, and the utilization of their data centers. Their financial performance lies on the prices customers are willing to pay relative to those costs.
Investors are now assessing whether the industry’s enormous capital commitments can be supported by durable demand rather than expectations of future growth alone. Earlier this week, Broadcom Chief Executive Hock Tan compared investors’ rush into anything associated with AI to “casino gambling,” reflecting concerns about speculative behavior across the sector.
Market movements on Thursday added to that unease. CoreWeave shares fell another 2% after plunging 15% on Wednesday, leaving the company with a loss of nearly one-fifth of its market value over two sessions. Its shares are now down 10% since the beginning of the year.
By contrast, shares in rival Nebius have risen approximately 114% over the same period. Nebius has no publicly traded CDS, preventing a direct comparison of its credit-insurance costs with those of CoreWeave.
The contrasting share-price performances show that investors are not treating all AI infrastructure businesses equally. However, equity returns alone do not establish which company has the stronger balance sheet or more sustainable business model.
Debt Becomes A Critical Test for The AI Trade
The latest surge in credit protection costs introduces a more consequential test for the AI boom, questioning companies’ ability to convert rising demand into sufficient cash flow to support the debt used to finance expansion.
Until recently, much of the investment debate focused on whether AI companies could grow quickly enough to justify their valuations. The deterioration in CDS markets suggests that financing costs and balance-sheet resilience are becoming equally important considerations.
Higher borrowing costs do not necessarily derail AI investment. Companies with strong cash generation, disciplined capital allocation, and reliable customer demand may be able to sustain their spending. The risks are greater for businesses that depend on continued access to debt markets while operating losses and infrastructure commitments remain substantial.
For investors, CoreWeave’s credit-market distress is seen as more than a decline in the shares of one company. It is a warning that the economics of AI infrastructure must be assessed alongside the technology’s growth prospects.
If borrowing costs remain elevated and returns take longer to emerge, highly leveraged operators could face pressure to slow investment, seek additional financing, or renegotiate their financial commitments. Such adjustments could affect suppliers and other companies linked to the data-center investment cycle.
The immediate concern is not that demand for AI has disappeared. Anthropic’s reported revenue growth points in the opposite direction. Rather, the market is beginning to question whether commercial demand across the industry will generate returns quickly enough to support the scale and financing costs of the infrastructure being built.