Invest1 publisher3 min readPublished
Anthropic's reported November IPO date is a bet on one audited quarter
Anthropic has reportedly moved its roughly $2 trillion IPO from October to November, a delay one investor ties to audited Q3 numbers. The extra month gets audited figures on its revenue race with OpenAI into the filing, and keeps the listing exposed to a window that can shut without warning.
The Investor · Invest desk
What happened
- Anthropic has reportedly moved its IPO, at a valuation of around $2 trillion, from October to November, according to the latest 20VC x SaaStr episode.
- Harry Stebbings countered that a company expecting 20x to 30x oversubscription normally does not wait, and that the delay could reflect stress in pre-marketing.
- Reports cited on the episode put OpenAI's cash burn at $278 billion through 2030, with its cash running out around 2028.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- exposure A November prospectus sets Anthropic's audited third quarter against the quarter OpenAI used to fight back, so any slippage in that race becomes an audited public record.
- contradiction Two hosts on the same episode drew opposite signals from one slip, so the date alone cannot tell investors whether demand for the deal is strong or strained.
- precedent A listing near $2 trillion would give OpenAI's investors a public comparable for its rumored $1.5 trillion round while OpenAI still has a gap of about $156 billion to fund.
O'Driscoll's case depends on one quarter against one rival. By his account, Anthropic had a huge second quarter and passed OpenAI in revenue [2], and OpenAI hit back hard in July with its own third-quarter story [3]. Listing in October would mean selling stock on a quarter OpenAI has already answered, with the newer quarter closed but unaudited [3][4]. "If we do this in October, it's going to be a lot of explaining. If we do this in November, the numbers will talk," O'Driscoll said [5].
The episode noted that some took the move as a crack in the market [17]. Stebbings's objection is a narrower version of that: stress in pre-marketing talks [6]. His objection and O'Driscoll's explanation can both be true. If pre-marketing investors pressed on the third quarter, an audited version of it is the answer a banker would reach for, and a one-month slide gets it. The hosts settled on the strength reading, describing the wait as what a company does when it thinks it is in a strong position with time on its side [15].
O'Driscoll would call the wait a good decision 90% of the time [7]. In the other 10%, the window shuts, and, according to the episode, the company is left regretting that it did not take the $100 billion [7]. At face value, a one-in-ten chance of losing $100 billion is an expected cost of $10 billion for a one-month slip [2]. Against the reported $2 trillion valuation, $10 billion is half a percent [3]. An audited quarter that lifts the price by more than half a percent covers it.
The quarter matters because of the growth OpenAI is forecasting. O'Driscoll put OpenAI at about $35 billion of annual recurring revenue at the end of this year and about $350 billion in three to four years [11]. Ten times in three years is about 115% growth a year; over four, about 78% [4]. He called that close to modest, since Anthropic did 10x in one year [12].
Jason Lemkin doubted the cost side. "I bet it's more," he said of the $278 billion burn forecast, adding that burn at the fast-growing companies he has backed came in 30% to 50% over plan [16]. Apply those overruns and the figure becomes $361 billion to $417 billion [5]. According to O'Driscoll, the burn stays that low only because Oracle, Nvidia and others build the data centers behind roughly $700 billion of capex and lease the capacity back [13].
The episode covers one listing and does not describe other AI companies preparing to go public. On this record, November is one company's decision about one quarter, and it does not show a timetable for the rest of the sector. I think O'Driscoll has the better argument. A one-month slide matches the gap he describes between a quarter's close and its audited numbers [4]. What Anthropic gives up is an October listing built on a second quarter that OpenAI has already answered [3]. The view fails if the date slips past November, or if the audited third quarter shows Anthropic behind OpenAI; either result would favor Stebbings [6].
What to watch
- Whether the offering prices near the reported $2 trillion valuation; a range well below it would support Stebbings's pre-marketing-stress reading.
- Whether OpenAI closes its rumored round at $1.5 trillion before or after Anthropic's shares start trading.
- How the prospectus discloses product liability for autonomous agents, given O'Driscoll's argument that a $2 trillion company can self-insure without Munich Re.