CoreWeave had a bad week last week. Its bonds slipped into distressed territory by a common definition — when investors demand 1,000 basis points more in yield than they’d get from government bonds. Last Tuesady Chamath “SPAC King” Palihapitiya noticed the distressed debt, and also asked whether the AI buildout was a private credit bubble. Its shares are down more than 20 percent since Tuesday’s close, which is kind of a lot! There even seemed to be some collateral damage in the market on other AI companies, too.
Now, I have also asked if the AI industry is a bubble, but when the SPAC King echoes concerns held by a lady whose ongoing question about AI is “How are we planning on affording this, champs?”, that’s a very different signal. When I say “yikes,” it’s one thing. When Elon Musk says it, that’s something else.
When you talk about wanting to tap into the enormous sums of money that insurers have, that’s the kind of thing that makes investors afraid that CoreWeave is going to borrow even more
To make things worse for CoreWeave, Brannin McBee, one of the company’s cofounders, told the Financial Times — also last week, on the same day as Palihapitiya’s post — that the company wants to have its debt be investment-grade, rather than junk, so that it can borrow from insurers, as their capital is, in his words, “unlimited.” This is not entirely unreasonable — earlier this year, CoreWeave secured its first investment-grade chip loan. But when you talk about wanting to tap into the enormous sums of money that insurers have, that’s the kind of thing that makes investors afraid that CoreWeave — already a massive debtor — is going to borrow even more.
“The problem is not that credit profiles have changed in the last year, it’s that we’ve got another year and a half of supply coming,” Bloomberg quoted John Yovanovic, co-head of leveraged finance at MetLife Investment Management, as saying.
I have seen people suggest that this means the AI bubble is bursting. I am not so sure. A plunging stock price is not necessarily financial doom. Stocks are an expression of confidence in a company’s future, not a reflection of day-to-day business. But CoreWeave has a lot of very expensive debt, and if that debt gets more expensive, that makes it even harder for CoreWeave to turn a profit.
CoreWeave is the poster child for AI debt, and as of its second quarter earnings, owed more than $35 billion. That’s partly because it pioneered GPUs as collateral for loans, and partly because CoreWeave doesn’t have a real business beyond selling compute, a business on which it does not make money — the company lost more than a billion dollars in the first half of the year.
The anxieties investors have around CoreWeave are, in fact, present for all neoclouds
So why am I not ready to do a victory lap on this? After all, I have suggested that this business is long-term unsustainable. The answer is “remaining performance obligations” from its customers — $103.7 billion of them as of the second quarter. About 40 percent of that money is due to CoreWeave in 2028, and the companies who’ve agreed to these contracts can’t cancel them except under some very limited circumstances. Its biggest customers are Microsoft, OpenAI, and what is probably Nvidia, though CoreWeave doesn’t disclose who its third-biggest customer is. Those three generated almost three-quarters of the company’s second quarter revenue.
Microsoft and Nvidia are probably good for the money — at least, as long as it benefits them to be good for the money. (I suspect they realize that if CoreWeave goes belly-up, there will be plenty of collateral damage in the market, so they are unlikely to push the company over the edge themselves.) OpenAI is shakier, but unless OpenAI is in significantly more financial distress than we all realize, CoreWeave is probably fine for a while. As long as it can continue to pay off its debt, the music keeps going. Only when the AI boom falters are we going to find out if it’s actually a viable business.
Still, the angst around CoreWeave has made things shaky for other companies that have taken out significant debt in the AI buildout — Oracle’s shares, for instance, which have fallen almost 10 percent in the same timeframe. There’s been a sell-off in Oracle bonds, too. The anxieties investors have around CoreWeave are, in fact, present for all neoclouds. It’s just not clear the bare-metal business is a good one.
Anyway, some people are buying the dip.