Why Anthropic’s Rivals are Rooting for its IPO

Anthropic needs its IPO to go well. So do its rivals.

The AI model developer’s listing, likely to take place in a matter of weeks, will say as much about access to capital for everyone else as it does about Anthropic’s viability as a business.

There is little doubt Anthropic will be able to raise the up to $100 billion it is seeking. Whether competitors like Meta Platforms or OpenAI will ride those coattails to raise gobs of their own capital is another question.

Equity investors have already piled into SpaceX’s listing this summer, absorbed a nearly $85 billion equity raise by Google parent Alphabet and are currently being plied by Meta Platforms for more cash. The latter wants to raise nearly $100 billion, it said Monday, mostly by issuing new shares into the market over time.

The waters already look less inviting. AI-linked stocks have fallen pretty much universally over the past month, and so-called neoclouds like CoreWeave and Nebius have plunged. CoreWeave is down more than 35% in that period. Meta’s stock also fell after it announced its equity-raising plans.

Signs of skittishness are popping up on the debt side, too. Big tech companies are expected to sell around $250 billion of bonds this year, followed by around $400 billion next year. That is a lot of issuance in a higher interest-rate environment.

Already-issued debt isn’t looking terribly healthy, either. CoreWeave’s unsecured corporate bond prices have cratered recently. That reflects uncertainty about its debt-fueled business model, but also about AI writ large.

Banks have been struggling to sell debt instruments backed by AI-related loans. The AI computing player Broadcom had to sweeten its guarantees and work out higher interest rates to get banks to move forward this week on a $63 billion deal to finance chips for Anthropic.

The scramble for funding underscores how the AI craze has shifted to a phase where capital is more constrained. And access to it at a reasonable cost is becoming more of a deciding factor in the race than ever. Winning may still be about technology and computing power, but finding money to buy enough of it has gotten significantly harder.

That is why Meta in particular could use a strong Anthropic IPO.

Mark Zuckerberg’s company is raising $37.5 billion in equity as a first step in its broader share-issuance plan, anchored by investments from AI-chip leaders Nvidia, Advanced Micro Devices and Broadcom. But analysts estimate the company’s capital spending next year at around $200 billion, meaning it needs the capital spigots to stay open.

OpenAI, Anthropic’s nearest competitor, badly needs the IPO to work for slightly different reasons. The AI lab that touched off the boom almost four years ago is vying for its own IPO next year.

Any sign of investor trepidation about Anthropic could put that plan—and Sam Altman’s hopes of scaling up his company’s computing power to stay in the frontier-model mix—on ice.

If you’re worried about circular financing propping up AI, this news won’t bring much comfort: Broadcom and a consortium of banks are getting together to fund an unprecedented $63 billion chip deal with Anthropic.

The terms of the deal are complex, but it boils down to this: Broadcom and Google are designing the chips, which will be manufactured by Taiwan Semiconductor Manufacturing Co. A special-purpose vehicle, funded by the banks and supported by guarantees from Broadcom, then buys the chips and associated computing equipment. That special-purpose vehicle then leases the chips to Anthropic, and Anthropic’s payments service the debt.

That works as long as AI demand is strong and Anthropic continues to grow. Hiccups could arise, though, including business struggles for Anthropic or the banks having trouble selling the debt on to investors.

Percentage growth in Anthropic’s revenue in the first six months of this year compared with a year earlier, as disclosed in its IPO prospectus.

Meta Platforms used to be a company that didn’t dabble much in debt. AI has changed that.

In a stark illustration of how tech business models have changed since the boom, Meta’s long-term debt has skyrocketed, going from under $20 billion less than three years ago to more than $80 billion today. That is no surprise given how much money the company is plowing into AI. But it is a major switch for a company where the future long hinged on user growth and an advertising business built around it.

Meta hopes its big investments will pay off, and it has a hit with its Muse AI agent. Muse, though, likely won’t be enough to justify all the capital spending—and debt—that has transformed the company’s financial makeup.

Microsoft is making buyout offers to 60,000 employees, the latest bid by a big tech company to pare its workforce and make way for AI. Buyouts, as opposed to layoffs, can be tricky, though. In-demand employees may take them because they know they can easily find work elsewhere, leading to unwanted departures. But buyouts can be less damaging to morale than layoffs because of their voluntary nature.

WSJ AI & Business is a weekly look at AI’s transformation of the business world. This newsletter was curated and edited by Asa Fitch. Reach him at asa.fitch@wsj.com (if you’re reading this in your inbox, you can just hit reply). Got a tip for us? Here’s how to submit.