Anthropic IPO targets November debut despite AI scrutiny

Anthropic IPO preparations point to a possible November listing, people briefed on the discussions said, testing AI demand near $2 trillion.

The Claude developer could begin formal investor marketing during the week of Nov. 9, putting a first trading day before the Thanksgiving holiday on Nov. 26, those people said. They said the company is still expected to list by year-end, although the timetable remains under discussion and could move again.

A Thanksgiving cutoff shapes timing

The calendar matters because equity issuance usually slows around the holiday, leaving a narrow window for a large technology offering. Anthropic had previously been positioned to file publicly after the summer, according to people briefed on the earlier planning, before the schedule slipped.

The company is also set to meet prospective investors on Oct. 14 at its San Francisco headquarters, the people said. That meeting would give buyers an early look at the growth story, the loss profile and the governance questions attached to one of the most closely watched AI labs.

Anthropic has risen with demand for frontier AI models, but the offering work is arriving as OpenAI has gained sales momentum in recent months, according to people tracking the companies. OpenAI has postponed its own IPO planning, with CEO Sam Altman saying a public listing would be ill advised for now.

Claude economics face scrutiny

Investors considering Anthropic are being asked to value a company with fast revenue growth and heavy losses. People briefed on the discussions said some prospective buyers see a fair valuation in the $1.8 trillion to $2 trillion range, a level that would place the offering among the largest technology listings ever attempted.

The documents cited by people briefed on Anthropic's finances put its 2025 net loss at almost $42 billion, roughly five times the about $8.3 billion loss recorded a year earlier. Full-year revenue was roughly $4.6 billion in 2025, up from $386 million the prior year, while the operating loss widened to more than $8 billion.

Most of the net loss came from accounting movements rather than day-to-day operations, according to those documents. A change in the fair value of liabilities accounted for more than $34 billion of annual losses, a figure that will require careful explanation if the company files public IPO documents.

AI safety is another variable in the deal. People briefed on the matter said public scrutiny has increased after high-profile hacking incidents involving rogue agents, while Anthropic Chief Executive Officer Dario Amodei has argued on his personal website that the pace of new model development should slow.

New listings lag indexes

The IPO market is not giving Anthropic a clean backdrop. Oura recently became the third company in a matter of weeks to postpone a planned first-time share sale, after some potential buyers resisted a fully diluted valuation target of about $15 billion.

Performance data for this year's new listings also shows why bankers are watching demand closely. Excluding record-setting listings by SpaceX and SK Hynix, more than 100 newly listed stocks had a weighted-average loss of 4%, compared with a 12% gain for the S&P 500 Index and a 20% rise for the Nasdaq 100 this year, compiled market data showed.

If Anthropic starts marketing in the week of Nov. 9 and prices before the holiday, the deal could reset investor appetite for AI infrastructure and model companies by creating a fresh public benchmark. That outcome would support Anthropic's access to capital, give peers a valuation reference and feed a broader macro trade tied to compute spending, data centers and semiconductor demand.

If the schedule slips again, the mechanism cuts the other way: investors would have more time to press on losses, safety risk and the durability of AI revenue growth. For Anthropic, that could mean a narrower valuation range; for the sector, it would add another delay to a listing market already trailing major equity indexes.