Here’s What CoreWeave’s Credit Woes Tell Investors About the AI Trade

Key Takeaways

• The value of debt issued by CoreWeave, an AI-linked neocloud firm, has plummeted as investors become more discerning about the AI infrastructure buildout.
• Pure-play AI bets like CoreWeave carry additional risks because their outlooks are tied to a new, uncertain industry.
• Companies are issuing more AI-related debt, but analysts don’t see cracks in the broader credit market.

The artificial intelligence trade is cracking, with shares of related firms looking wobbly as investors grapple with doubts about the sustainability of the massive debt load companies are accruing to fund the AI buildout.

Case in point is CoreWeave CRVW, a “neocloud” company that leases out computing power in the form of chips and other physical infrastructure necessary to run AI models. After a highly anticipated public debut in early 2025, the firm’s shares have lost more than half their value.

At the same time, CoreWeave’s bonds have plummeted in value, with some trading in territory Wall Street would consider distressed.

CoreWeave isn’t the only neocloud firm to run into trouble. Shares of Nebius NBIS have slipped, Elice Group had a dismal IPO showing in South Korea, and Australia’s Firmus terminated a planned public listing due to concerns about its lofty valuation. CoreWeave is the highest-profile case for US investors. It’s also a bellwether for an existential issue facing Wall Street: How much AI borrowing can the market sustain?

As investors grapple with that question, pure-play AI firms like neoclouds look especially vulnerable, according to Sameer Samana, head of global equities and real assets at the Wells Fargo Investment Institute. Unlike megacap hyperscalers, neoclouds depend totally on the AI buildout. “Everything has to go right,” he says. “You live by the sword, die by the sword.”

CoreWeave’s Debt Drama

At the center of CoreWeave’s story is a growing debt burden, both on its balance sheet and in the form of special purpose vehicles, which fund specific projects. Morningstar analysts Luke Yang and Javier Correonero expect CoreWeave’s total debt load to surpass $100 billion by 2030. They say its aggressive financing makes it one of the most leveraged companies under Morningstar’s technology coverage.

As investors grow more discerning about AI borrowing, some of CoreWeave’s debt has plummeted in value. For example, the yield on a 9.625% senior note due in 2032 climbed above 15% earlier this month, while the bond’s price dropped to less than 80 cents on the dollar. That’s a sign investors are demanding a sky-high premium to protect against the risk of default.

Another gauge of perceived credit risk—spreads on CoreWeave’s credit default swaps—tells the same story. The wider the spread, the more expensive it is to insure against a debt default. CoreWeave’s five-year swap spreads are now around 1,100 basis points, indicating buyers are demanding a 10% premium (or more) to take on the risky loan. That’s roughly double the near-5% premium in June.

That has implications in the equity market, too. “High leverage nevertheless makes CoreWeave a risky bet for equity investors,” Morningstar’s analysts write. “If long-term demand for AI undershoots industry stakeholders’ expectations and CoreWeave sees material changes to its financing terms, its equity value can quickly go to zero.”

CoreWeave’s Bottom Line

Despite its ballooning liabilities, Morningstar analysts have confidence in CoreWeave’s business model. They expect it to benefit from a close relationship with Nvidia NVDA, plus a stream of prepayments on its contracts over the next few years.

Morningstar DBRS assigns CoreWeave a BB rating, below investment grade. But that doesn’t necessarily mean red flags in the near term.

Victor Leung, senior vice president and sector lead for corporate ratings, asset finance at Morningstar DBRS, says CoreWeave is a strong player among neocloud firms. Relatively speaking, he says, CoreWeave is better capitalized, with more experience and a better track record. He says the risk investors perceive is larger than one company. CoreWeave is “caught up in the AI story ... They’re operating in the AI industry, which has this uncertain question mark” of whether the rate of growth and borrowing is sustainable.

At the same time, the uncertainty is playing out in real time for CoreWeave. Leung notes that hyperscaler customers like Oracle ORCL, Broadcom AVGO, Meta Platforms META, and others are backing away from risk in longer-term contracts. Leung says that could make it harder for CoreWeave to close future financing deals, or that those deals will carry lower ratings or higher pricing to account for the additional risk the firm must assume.

AI Financing Under the Microscope

Overall, analysts say recent AI-related wobbles in credit markets are not a sign of weak industry fundamentals.

Lotfi Karoui, a multi-asset strategist at Pimco, argues they’re better viewed as an indication that markets are repricing some debt to account for higher interest rates, more debt in the market, and higher-quality alternatives. “The next financing wave will likely be larger and more varied, forcing lower-quality borrowers to compete with hyperscalers, investment-grade-tenanted data centers, and better-protected asset-level structures,” he writes. “With all-in yields already elevated, investors have less reason to reach for incremental credit risk.”

Rebecca Venter, senior fixed income client portfolio manager at Vanguard, says markets are heading for a “show-me” period, when investors will focus on the results of the AI infrastructure buildout, especially at hyperscaler firms. “The market will need to see profitability and actual monetization of the technology,” she says. She expects more clarity about those results to hit corporate bottom lines by the second quarter of next year.

An AI Contagion in the Credit Market?

Zooming out, analysts say that while uncertainty around AI issuance is translating to somewhat wider credit spreads, the credit market as a whole is not coming under stress. AI-related debt remains a small portion of the corporate bond market, and Samana of Wells Fargo points out that investment-grade credit spreads have widened only slightly, remaining well below last year’s levels. “There’s no broader contagion” from AI worries in the wider credit markets, he says, and capital markets remain “wide open” for the right companies. Meta’s recent announcement of a $95 billion stock sale to fund more investment in AI infrastructure is evidence.

Vanguard’s Venter says the credit market’s fundamentals still look good, and she expects spreads to remain in a tight range. That said, “the risks of more AI issuance coming to market and the market wanting to see more in terms of profitability are definitely there.” All that borrowing from both companies and governments could also put upward pressure on interest rates.

What’s Ahead for the AI Trade

Wells Fargo’s Samana says further instability in the AI trade could prompt investors to consider exit strategies, and pure plays like neoclouds could be first out the door. If it all goes south, “you want to favor large over small ... and probably favor the check receivers over the check writers,” he says. “You want to be somewhat cognizant of valuations, of whether companies can self-fund or rely on capital markets.”