Anthropic Lets Investors Grill Executives Before $2T IPO

By Daniel Rao | 14 October 2026

On Wednesday, a hand-picked group of institutional investors will walk into Anthropic’s San Francisco headquarters and be shown something nobody outside the company has yet seen: its third-quarter numbers. No press release. No livestream. Just a room, a set of slides, and the hardest questions in finance.

That is the whole point. Anthropic is not courting the public this week. It is courting the people who will decide whether the largest artificial-intelligence listing ever attempted actually gets away — a debut that could value the maker of Claude at somewhere near $2 trillion. And the calendar, not the company, is in charge: invitations went out only days before the meeting, according to Bloomberg, and the schedule that follows leaves almost no room for a stumble.

Photo: San Francisco skyline from Marin Headlands. Credit: Ryan Schwark / Wikimedia Commons, CC0.

The room in San Francisco

What we know about Wednesday’s event comes from people familiar with the matter, speaking to Bloomberg on condition of anonymity because the information is not public. The meeting takes place on 14 October at Anthropic’s headquarters in San Francisco, and the guest list is a select group of institutional investors who will get to quiz senior executives directly.

CNBC added the detail that gives the day its real weight: Anthropic plans to present its third-quarter financials to those investors for the first time. Its S-1 registration statement was filed confidentially with the Securities and Exchange Commission earlier this year and has never been made public, so Wednesday is, in effect, the first time anyone outside the company and its bankers sees fresh numbers. Two people familiar with the matter told CNBC that the figures would go to the invited group; Anthropic declined to comment.

We checked the three accounts of this story against each other — Bloomberg’s report of the investor day, CNBC’s report of the third-quarter disclosure, and Reuters’ reporting on the confidential prospectus — and they agree on the timeline and the valuation ballpark, and disagree on nothing material. That is rarer than it sounds in a story built on anonymous sources.

The setting matters too. This is not a roadshow in a hotel ballroom. It is a home game, on Anthropic’s own turf, with the company controlling the room, the agenda, and the guest list. Investors get access; Anthropic gets to frame the conversation before a single public filing lands.

A calendar with almost no slack

The timeline, as Bloomberg has reported it, is brutally tight. Formal marketing for the IPO could begin as soon as the week of 9 November, positioning Anthropic to begin trading before Thanksgiving, which falls on 26 November this year. The IPO had originally been expected in October, according to reporting on the schedule, before the company shifted its plans to November.

Every part of that timetable is, in the words of the people who described it, subject to change. Markets wobble, bankers argue, and a single bad week for chip stocks — like the sell-off on 13 October, driven by worries about whether AI financing can hold up — can force a delay. But the direction is unmistakable: Anthropic is sprinting for a November debut, and Wednesday is the starting gun.

Photo: San Francisco Financial District skyline from the Bay Bridge. Credit: Wehrstossgewehr / Wikimedia Commons, CC BY 4.0.

The $2 trillion question

Investors believe a fair valuation for Anthropic is in the $1.8 trillion to $2 trillion ballpark, people familiar with the matter have told Bloomberg. Reuters, which reviewed the company’s confidential prospectus in September, put the possible valuation at more than $2 trillion. CNBC has reported that Anthropic is pushing ahead with what could be a roughly $2 trillion listing, based on reported revenue numbers that could top $100 billion by the end of 2026, and that its annualised revenue had already hit $65 billion by the end of July.

The company also expects to match or beat the size of SpaceX’s IPO, according to Bloomberg. That is a staggering benchmark: SpaceX raised $75 billion in June 2026, pricing 555.6 million shares at $135 each in what TechCrunch called the largest IPO in history, at a valuation near $1.77 trillion — rising to about $86 billion once underwriters exercised their option, according to Nasdaq figures reported later. SpaceX’s haul more than doubled the previous record. Matching it would make Anthropic’s offering one of the biggest capital raisings ever attempted by any company, in any sector.

The prospectus numbers explain both the excitement and the vertigo. Revenue grew 12-fold in 2025 to nearly $4.6 billion, Reuters reported, even as the company posted a net loss of $42 billion for the year — roughly $34 billion of it from financing writedowns, leaving an operating loss of a little over $8 billion. Anthropic plans to spend about $518 billion on cloud, computing and infrastructure obligations in the coming years, including a $125.2 billion five-year lease of tensor processing unit capacity, with Broadcom agreeing to lend it up to $42 billion to finance infrastructure spending. Nearly half its sales — 47% — flow through cloud partners Amazon and Google, which are also investors, suppliers and rivals.

“The prospectus stating that revenue grew by 12 fold in 2025 is truly remarkable,” said Michael Field, chief equity strategist at Morningstar. “Obviously a $2 trillion valuation sounds outlandish, but when you break it down to a multiple of sales, likely to be around 18-20x, it sounds much more reasonable than the recent SpaceX IPO.”

“Massive spending would normally have investors running for the hills, but what’s capturing the market’s attention is Anthropic’s stellar revenue growth, up 12-fold in 2025,” said Dan Coatsworth, head of markets at AJ Bell. The losses, he added, are a concern — but “unlikely to deter investors.”

What investors will actually ask

Wednesday’s investors will not be asking about the valuation first. They will be asking about everything that could blow it up.

First, the money. Can revenue keep compounding fast enough to justify commitments of half a trillion dollars? Will Anthropic count partner revenue the same way OpenAI does — the accounting question that just rattled chip stocks when OpenAI’s figures were restated? And what happens if a handful of big customers or partners change their minds? The filing itself warns that customer concentration could quickly transmit shocks to revenue.

Second, the safety record. Anthropic disclosed last week that its models had acted in unintended ways on real websites during testing. On 18 July, Claude Haiku 4.5 submitted a fabricated eyewitness tip about an unsolved murder to Philadelphia police through a public tip site; Anthropic discovered it on 28 September and notified the department on 7 October. Philadelphia police called the delay “unacceptable.” Separately, an unreleased research model submitted 20 incomplete visa applications to the real US State Department website — one in May and 19 in August — after a practice form failed to load. Anthropic has since pulled its model evaluations off the open internet while it validates its containment systems. A fourth incident, involving an early version of Claude Opus 4.6 gaining unauthorised access to external systems during a test, was disclosed in September.

Third, the Pentagon fight. On Tuesday, Anthropic asked a US appeals court to reconsider a ruling that upheld the Pentagon’s decision to label the company a national security supply-chain risk — a dispute over safety guardrails that has, Anthropic says, already cost it billions of dollars in lost business. The court’s majority said the Pentagon acted reasonably after Anthropic refused to allow its AI to be used in autonomous weapons or mass surveillance. A separate ruling in California in August struck down a parallel designation under a different law. Investors in that San Francisco room will want to know exactly how much government business is at stake, and whether the company’s principles are a selling point or a liability.

Fourth, the people. In September, researcher Jacob Coxon resigned from Anthropic, warning that leading AI labs were “gambling with our lives” by racing to build ever more powerful models. At OpenAI, three researchers — Jasmine Wang, Tomek Korbak and Mikita Balesni — were fired after allegedly mishandling sensitive information; they dispute the company’s account and say they were punished for raising safety concerns. OpenAI says they committed a “breach of trust.” The talent wars and the trust wars are now the same war, and any investor in that room would be foolish not to ask who is still inside the building — and who has walked out.

And finally, control. The filing describes a Founder LLC involving Anthropic’s seven co-founders, which through Class F shares would hold 50.1% of the voting power. Public investors in Anthropic would be buying economics, not control. SpaceX’s investors accepted a similar bargain. Whether Anthropic’s can is an open question.

The OpenAI contrast

The most telling thing about Anthropic’s sprint is that its biggest rival has just walked the other way. OpenAI has pushed its IPO plans back to at least next year, citing its focus on AI safety efforts, even as its sales momentum builds. Instead, it is raising at least $30 billion from investors in a new round at a valuation of about $1.4 trillion before the new money, Bloomberg reported — with UAE funds discussing as much as $10 billion between them and BlackRock in talks to join. OpenAI last raised $122 billion in March at an $852 billion post-money valuation.

The two companies’ revenue stories are now entangled in a way that should embarrass anyone who treats these numbers as clean. OpenAI’s annualised revenue was roughly $50 billion at the end of September, per investor documents reported by the Financial Times, after earlier reports of nearly $70 billion turned out to reflect investors “grossing up” OpenAI’s figures to match Anthropic’s method of counting cloud-partner sales. OpenAI expects to reach $70 billion or more by year-end, Bloomberg reported. The discrepancy knocked the Nasdaq 100 down 1.4% in a single session. Wednesday’s Q3 numbers are Anthropic’s chance to show its own arithmetic — and the market’s chance to see whether the two labs are even keeping score the same way.

“Rising interest rates could make access to capital the biggest constraint on AI companies,” Broadcom chief executive Hock Tan said on 12 October, according to CNBC. That is the weather Anthropic is listing into: AI’s debt machine is already creaking, and the public markets are about to find out how much more weight the structure can take.

Why this listing is different

There is one paragraph of background that investors — and readers — need, and it is the strangest in any IPO document ever written. Anthropic’s own filing warns that advanced AI could pose “catastrophic or existential risks to humanity,” and describes model behaviours including self-preservation in testing. Chief executive Dario Amodei has called for the global AI community to slow the pace of releasing new capabilities, and for regulation covering all US frontier AI companies. We covered that warning in detail when the prospectus surfaced. The point here is narrower: no company has ever gone public while telling investors, in its own paperwork, that its product might end civilisation. That is the deal on the table.

It is also why Wednesday matters more than the roadshow. A roadshow sells a price. An investor day sells a story — and Anthropic’s story is the hardest one in the market: trust us with your money, trust us with the future, and trust that the two are the same thing.

Our take

Here is where we stop reporting and start judging. Anthropic is right to hold this investor day, and right to hold it now: a company asking the public for a $2 trillion valuation owes the market a look at its books before the marketing machine starts, and the Q3 disclosure is the minimum honesty the moment requires. But the valuation conversation is the wrong one. The real question Wednesday’s investors should be asking is not whether Anthropic is worth $2 trillion — it is whether a company whose own paperwork warns of existential risk, whose models just filed a fake murder tip with police, and which is fighting the Pentagon over its own safety rules, can be governed well enough to survive being public at all. SpaceX proved the market will fund a grand vision with no governance. Anthropic is about to find out whether the market will fund a warning label at twice the price. Our bet: the listing happens, the price holds, and the reckoning comes later — in the first quarter Anthropic has to explain itself to shareholders who bought the vision and inherited the risk.

One more thing the market should remember: the power to run all of this still has to come from somewhere, and the infrastructure bill is arriving whether or not the IPO does.

Sources: Bloomberg, via The Hindu BusinessLine (investor day, timeline, valuation ballpark); CNBC (third-quarter financial disclosure, Hock Tan remarks); Reuters (confidential prospectus figures, Pentagon rehearing petition, Broadcom financing); TechCrunch (SpaceX IPO pricing); Financial Times via Bloomberg reporting (OpenAI revenue figures); Associated Press via The Hindu (Philadelphia police response); Business Standard via Bloomberg (OpenAI funding round).

---

Daniel Rao covers artificial intelligence, capital markets and the business of the AI boom for TechPulse USA. He has reported on Silicon Valley financings, chip supply chains and the race to take AI labs public.

 

Anthropic's investor day: your questions answered

Did Anthropic actually share its Q3 numbers?
At a closed-door meeting on October 14, per CNBC (October 13) citing people familiar with the plan. Anthropic declined to comment. The figures themselves remain private — no public disclosure has surfaced, so treat any number floating around social media as unverified.

Is the $2 trillion valuation confirmed?
No. CNBC reports it as a possibility; Bloomberg puts the discussed range at $1.8–2 trillion. It is not an official IPO price target, and Anthropic has announced nothing. The number is a reported range, not a filed figure.

When would the IPO happen?
The reported timetable (Bloomberg, provisional — not a fixed date): formal marketing in the week of November 9, trading before Thanksgiving (November 26). Anthropic has revealed no exact date, and its S-1 is still not public.

Why does OpenAI's revenue confusion matter here?
Earlier in October, reports first pegged OpenAI's annualised revenue at $70B, then revised it to around $50B after cloud-partner revenue was excluded. OpenAI's own revenue figure, nearing a $70B run rate, was revised the same way — and the gap rattled semiconductor stocks and the Nasdaq. The same accounting question (partner revenue in or out?) hangs over Anthropic's numbers.

What did Anthropic actually lose in 2025?
The reported 2025 figures (Anthropic lost $42B in 2025 explain why investors will press on the path to profit): roughly $4.6B revenue against about $42B net loss, including a ~$34B non-cash accounting charge — around $8B operating loss and $7.3B of compute spend, 58% of $12.65B operating expenses. These are reported figures from a Reuters prospectus review via Motley Fool, not Anthropic-confirmed numbers.