AI giant Anthropic files for an IPO that could break records

Anthropic is taking a big step towards a blockbuster stock listing that would rank among the biggest IPOs ever. Patrick Sison/AP Photo hide caption

Patrick Sison/AP Photo

Sign up for alerts on breaking news and exclusive reporting from NPR

The artificial intelligence startup Anthropic, maker of the hugely popular Claude AI models, has taken a major step toward holding an initial public offering that analysts expect to be one of the biggest ever.

Anthropic filed final paperwork with regulators on Monday for an IPO under the ticker symbol ANT that reports have said could value the company at between $1.5 trillion to $2 trillion.

That would put it in league with Elon Musk's rocketry and AI company SpaceX, which was valued at around $1.8 trillion when it listed in June. Anthropic's listing prospectus did not say how many shares the company will offer or at what price range.

The filing of a public version of a registration statement called a Form S-1 with the Securities and Exchange Commission is a major milestone in Anthropic's meteoric rise. It was launched just five years ago by a breakaway group of OpenAI researchers, led by Dario Amodei, who were concerned that that company, run by Sam Altman, was not on a path to develop AI safely.

Anthropic has since become a juggernaut, and OpenAI's biggest competitor, with some of the most widely-used AI models on the market. But the company has faced controversy, including a conflict with the Pentagon over how its models can be used, and questions about Amodei's dire warnings on AI risks.

AI stocks have been hit by volatility 

The filing comes at a volatile time in the stock market for AI-related companies, as investors waver between exuberance about the positive effects AI could have on the economy and concern about safety and the tsunami of investment in the technology.

The IPO, which could happen as soon as November, will be one of the most eagerly anticipated public share offerings in years. OpenAI, which makes ChatGPT, is also planning an IPO, but those plans are temporarily on ice, with Altman ruling out a listing this year as "ill-advised" in the face of swirling AI safety concerns.

"Going public has a lot of advantages. It brings in capital for growth. It makes it easier for you to raise additional capital elsewhere. And it makes it possible for your existing investors to cash out," said Jim Angel, an associate professor and markets specialist at Georgetown University.

But, he added, "there's concern that we may be in an AI bubble. We're in a very dynamic industry, to say the least, and we can expect a lot of volatility in the markets." Still, Angel said, it makes sense for Anthropic to "strike while the iron is hot" and list shares now to get as high a valuation as possible.

With SpaceX already trading publicly, Anthropic's IPO could test investor appetite for AI businesses at a time when markets are jittery about whether or not they can be profitable enough to justify the mind-boggling sums of money being invested in AI. SpaceX's shares soared briefly after the listing, but have come back down to around $150, where they started trading.

Concerns remain about the AI sector

The balancing act for companies trying to ride the AI wave is real. The data center and infrastructure provider CoreWeave saw its share price plummet this month when investors got nervous about its ability to finance its expansion plans and its bond yields leapt. Other major players, including Intel, Meta and Nvidia, have also been buffeted by concerns and watched their share prices slide in recent weeks.

Debt financing has been a growing concern as AI-related companies scramble to raise money to meet demand at a time when interest rates are poised to keep rising as the Federal Reserve tries to douse stubborn inflation.

The S-1 filing that tees up the IPO flagged what Anthropic sees as the risks from AI — and to its business model. Among them are the risk that its AI models could develop faster than the company can keep them safe, or that they could be used to cause harm — or take harmful actions on their own. It also flagged the risk of heightened scrutiny or political backlash for the company's views on policy and regulation.

Amodei himself has issued some of the AI world's sharpest warnings about the possibility of AI spiraling out of control and causing harm, and urged AI companies to "pace the frontier," or to slow the development of their most capable models.

Anthropic has also clashed with the Pentagon this year over how the military can use its models. Amodei has drawn a line in the sand when it comes to using Anthropic's AI for autonomous weapons and mass surveillance of Americans.

In response, the Pentagon blacklisted Anthropic and designated it a "supply chain risk" — a label previously reserved for entities affiliated with U.S. adversaries. Anthropic has sued the Pentagon in two federal district courts. Those legal battles remain ongoing.