Anthropic has moved its listing from October to November, and almost every account of the delay has framed it the same flattering way: the company wants one more set of results to show investors before it prices. That framing treats a postponement as a victory lap. Read the same fact from the other side and it says something harder. A company that could justify a $2 trillion valuation on the numbers already in hand would not need to wait for the next set.
The Wall Street Journal reported on September 18 that the timetable had slipped, with the public S-1 now expected in late September and the investor roadshow no earlier than mid-October. The target has not changed: a valuation of as much as $2 trillion and a raise of up to $100 billion.
The Gap the Delay Has to Close
The number that matters here is not $2 trillion. It is $965 billion, the mark Anthropic’s own investors set in May when the company closed a $65 billion Series H. Those were sophisticated buyers with access to the books, pricing the company four months ago. The public ask is roughly double that.
BTN has already reported that the $2 trillion figure originated with Anthropic’s investors rather than with Anthropic. The delay is the follow-through on that story. The people who set the number are not the people who now have to defend it in a roadshow, and the company has decided it would rather defend it with a September quarter in hand than without one.
To be clear about what those numbers look like, the growth is not in dispute and it is genuinely extraordinary:
- End of 2025: roughly $9 billion annualized
- February 2026: $14 billion
- April 2026: $30 billion
- Late May 2026: $47 billion
- End of July 2026: more than $65 billion
OpenAI’s comparable figure was about $40 billion as of August. Anthropic has passed its closest rival on revenue, and some analysts expect the run rate to clear $110 billion by year end. SemiAnalysis has projected the company’s first quarterly GAAP profit, near $1 billion, in the very quarter Anthropic now wants to show.
So the delay is rational. It is also an admission. If Q3 is the evidence that turns $965 billion into $2 trillion, then as of today that evidence does not exist in public, and the company knows a roadshow without it would be a harder sell.
What Waiting Costs
Delay is not free, and this is the part the coverage has skipped almost entirely.
First, the calendar. Pushing the listing into November puts pricing within days of the US midterm elections. That is a window institutional syndicate desks generally avoid for a deal of any size, and this would be the largest of all time. Anthropic is proposing to raise up to $100 billion against SpaceX’s record, which BTN covered when it became the largest IPO in financial history in June at $75 billion, rising to $86.2 billion once the overallotment was exercised. Saudi Aramco’s 2019 record was $25.6 billion. A deal of this scale needs a quiet tape, and the first week of November is not reliably quiet.
Second, the competitive position is not frozen while Anthropic waits. OpenAI’s GPT-6 Astra has been clawing back enterprise accounts, open-weight releases from Chinese labs and Meta are compressing what anyone can charge per token, and infrastructure costs keep climbing. Every one of those pressures shows up in a Q3 gross margin line. Waiting to show the quarter means waiting to show the margin, and the margin is where a price war lands first.
Our Read: Watch the Margin, Not the Run Rate
The delay is the most informative thing Anthropic has done since it filed confidentially in June, and it should be read as caution rather than swagger.
Our position is that the revenue story is real and the valuation ask is not yet supported. A company growing from $9 billion to $65 billion of run rate in nineteen months has earned an extraordinary multiple, and approaching GAAP profitability at that scale is rarer still. None of that is the question. The question is whether one more quarter can close a gap that informed private buyers priced at $965 billion in May, and the honest answer is that it depends entirely on what the margin line says, not the revenue line. Revenue growth at this pace is already assumed in the $965 billion. The incremental trillion is a bet on durable pricing power in a market where two of the three largest participants are actively cutting prices.
If Q3 shows margin expansion alongside the growth, the ask is arguable and the delay will look like good underwriting. If it shows revenue up and margin flat or compressing, then Anthropic will have spent six weeks of market goodwill to confirm the bear case, and the price will have to come down. Investors should watch the gross margin and the enterprise net revenue retention rate in the S-1, not the run-rate headline, which is the number everyone already agrees on.
There is also a governance point worth making. A valuation set by existing holders, defended in a roadshow by management, on evidence that arrives after the target is public, is an unusual sequence. It is not improper. It does put the burden of proof in an awkward place, and the S-1 due in the next fortnight is where that burden either gets discharged or does not.
The S-1 is the next real event, and it lands before the roadshow rather than after, so it will be read as the answer to all of this. Three lines in it will decide the deal: gross margin trajectory across 2026, the split between enterprise and consumer revenue, and the compute commitments Anthropic has signed. That third one has quietly become the sector’s biggest liability, and it will be the first thing a sceptical institutional buyer turns to.
Anthropic is not short of demand. It is short of one quarter of proof, and it has just told the market as much.