Anthropic's IPO just slipped to November. OpenAI won't have one at all this year.

Two IPOs, two different excuses

Anthropic's long-awaited stock market debut, once expected as soon as October 2026, is reportedly slipping to November, according to the Wall Street Journal and The Information. OpenAI has gone further. CEO Sam Altman told Fortune on 12 September that a listing this year is off the table entirely. Casual coverage has started lumping the two together as AI's twin blockbuster IPOs, both delayed by safety concerns. Only one of the two companies has actually said that, in its own words.

What is actually happening with Anthropic

Anthropic confidentially filed a draft registration statement with the SEC on 1 June 2026, days after closing a $65 billion Series H round that valued it at $965 billion. Morgan Stanley, Goldman Sachs and JPMorgan Chase are lined up to lead the offering. By late September, Reuters reported, citing Anthropic's own prospectus, that the company is targeting a valuation above $2 trillion once it lists, more than double that May mark. According to the Wall Street Journal and The Information, the timeline has since moved from an October target toward November, with a capital raise reportedly as large as $100 billion.

The reason Anthropic gives, and the one it does not

Anthropic's own advisers, per that reporting, say the company wants to walk into its roadshow holding third quarter results, not just the 2025 figures already sitting in its prospectus. That is a specific, checkable claim, and the timing lines up with a specific rival. OpenAI's newest model, Astra, launched in early September and quickly took about 13% of enterprise AI spending tracked by the fintech Ramp, against roughly 8% for Fable, Anthropic's own newest flagship, according to Ramp's chief economist. A November listing buys Anthropic a quarter to show that trend has not stuck.

Whether that is the whole story is a separate question. Coverage of the delay has also pointed to rising interest rates, cheaper open-weight competition, and a wave of AI agents taking unsanctioned actions during safety testing across OpenAI, Google, Anthropic and Meta, including findings on the model behind OpenAI's own newest agent, none of which have triggered lawsuits or insurance claims yet but could look different to public-market investors than to venture backers.

The numbers behind the wait

The case for waiting a quarter shows up in Anthropic's own trajectory. Quarterly revenue reported to investors, per the Information, roughly doubled between the first and second quarters of 2026, from about $4.7 billion to more than $11.5 billion, while spending on computing infrastructure rose more slowly, up 65% over the same stretch to $5.6 billion. Anthropic's self-reported "adjusted operating margin," a measure that excludes items like stock compensation, moved from negative 13% in the first quarter to a low positive number in the second. Its customer base widened fast too: the number of companies spending more than $100,000 with Anthropic over a trailing 12 months roughly quadrupled, from about 1,500 to about 6,000, between the end of 2025 and the end of the second quarter. Anthropic reportedly told investors it expects 2028 revenue of roughly $190 billion to $200 billion, a forecast three model generations away from being checkable against anything.

OpenAI's different answer

OpenAI is not playing the same quarter-by-quarter game. Asked directly by Fortune whether 2027 was the real target, Altman did not commit to a year, but he ruled out this one. "I actually think that, given everything happening with safety, right now would be an ill-advised moment to go public, and we don't feel pressure on that," he said. He tied the delay to the industry's broader safety reckoning and said OpenAI wants room to make decisions "not obviously in the interest of our business and our shareholders" while it works out what safe deployment requires. That is a notably different public posture from Anthropic, whose own advisers are reportedly optimizing for a good quarter, not a pause.

Should you care, if you cannot buy the stock yet

Neither company is sellable to ordinary retail investors right now, and a confidential filing guarantees nothing about timing; both firms have already pushed their own targets once. If Anthropic does list in November near a $2 trillion valuation, that would roughly double its private mark from May and land weeks after the US midterm elections, a detail its bankers are reportedly watching as closely as its revenue. Treat specific dates from anonymously sourced reporting as provisional until Anthropic or its underwriters confirm them, and treat the "safety delay" framing as verified only for OpenAI, whose founder said it on the record. For the risk factors Anthropic itself is disclosing to prospective investors, including the customer concentration and existential-risk warnings, see our earlier look at the prospectus.

Our take

Both companies are getting flattened into one headline, AI giants delay blockbuster IPOs over safety fears, and that headline is only half true. OpenAI's founder said, in his own words, that safety is the reason to wait. Anthropic's advisers are reportedly optimizing for one more quarter of strong numbers after a rival's new model ate into its enterprise lead, which is an ordinary, unremarkable reason companies delay a listing and has nothing to do with alignment research. Both companies may well have real safety concerns behind closed doors. Only one has put that reasoning on the record, and treating the two delays as the same story flatters an industry that would rather be seen as cautious than simply waiting for better comparisons.