The delay
Anthropic has moved its planned initial public offering from October to November, the Wall Street Journal reported on Thursday, citing people familiar with the matter. The company told Business Insider last week that it had picked Nasdaq and hoped to complete the listing in October; this is the first slip in that timetable.
The stated reason is straightforward. The extra weeks let Anthropic close its third quarter and put those figures in front of prospective investors before the roadshow — figures that will cover the period since OpenAI shipped GPT-6 Astra in early September. A company selling shares into a market that has just watched its largest competitor launch a flagship model would rather show the quarter than describe it.
Reports caution that the date may move again depending on market conditions.
The numbers behind it
The New York Times reported, and Bloomberg relayed, that Anthropic expects annualised revenue above $100bn this year. The trajectory that gets there is unusually steep: roughly $9bn at the end of 2025, about $65bn as a run rate in late July, and early backers projecting around $110bn annualised by the end of the year.
Those are figures reported by people close to the company rather than audited disclosures, and Anthropic has not published them itself. The prospectus will be the first document that has to stand behind any of it.
Investors discussing the deal have talked about a valuation around $2tn and proceeds of as much as $100bn, according to the same reporting. If it priced anywhere near that, it would be the largest listing on record, ahead of SpaceX’s in June.
The awkward part
The timing sits oddly against what Anthropic’s own chief executive has been saying. Dario Amodei has spent the past month publicly arguing that frontier labs should slow down together, and Anthropic named Accenture its first embedded safety evaluator on Thursday, committing at least $1bn over five years.
A company arguing for restraint while preparing the largest IPO in history has to make both cases at once: that the technology is dangerous enough to need pacing, and that its commercial trajectory justifies a two-trillion-dollar price.
Sam Altman took the other route. He ruled out an OpenAI listing this year, saying that given safety concerns “right now would be an ill-advised moment to go public”.
What to watch
The S-1, whenever it lands, is where the reported revenue figures either hold up or do not. It will also be the first time anyone outside Anthropic sees the cost side — what the company spends on compute, and how much of the revenue survives it. Until then, every number in circulation comes from people briefing reporters rather than from a filing.
