Disclosure. Anthropic is the company this article covers. Claude, the AI model used to help research and draft QuietWealth articles, is made by Anthropic. We have tried to hold this piece to the same standard as any other: figures are attributed to named sources, skeptics get as much space as supporters, and there are no price predictions or recommendations. Read it with that conflict in mind, and check the primary filing.

On October 26, 2026, Anthropic made its S-1 registration statement public, putting a proposed Nasdaq listing under the ticker ANT on the calendar. The filing is the first detailed look at the financial statements behind the fastest-growing company in the AI race, and it landed in the middle of a loud argument over whether AI spending is a boom or a bubble. This piece covers what the filing shows, what it leaves blank, and how analysts and investors are reading it.

Short answer: the filing shows extraordinary revenue growth, an operating loss that has nearly closed, a headline net loss of $80.7 billion that is mostly non-cash accounting, and very large compute commitments. It does not set a price, a share count, or a valuation. Reasonable people look at the same numbers and reach opposite conclusions, which is the heart of the "bubble" debate.

A note on sources. Core figures below were checked against the S-1 itself on SEC EDGAR, and we cite news coverage where the filing is silent or where reporters are describing expectations rather than filed numbers. Where outlets differ, we say so. Treat the filing as the authority, and verify before relying on any number here.

What has been filed, and what has not

The numbers in the filing, as reported

Anthropic S-1 figures (from the filing, with outlets noted where we relied on them)
ItemFigureReported by
Revenue, first half 2026$16.3 billion, vs. about $1.16 billion in H1 2025S-1, TechCrunch
Operating loss, first half 2026$693 million, vs. $2.1 billion in H1 2025TechCrunch
Net loss, first half 2026$80.7 billion, mostly noncash: a $72.4 billion fair-value charge on convertible instrumentsTechCrunch, The AI Insider
Adjusted operating income, H1$1.62 billion (10% margin)The AI Insider
Cash and short-term investments, June 30$131.65 billionThe AI Insider
Capital raised through June 30$141.1 billionIPOScoop
Customers spending $100K+ a year (Sept. 30, trailing 12 months)More than 11,900; more than 1,900 at $1M+; about 250 at $10M+TechCrunch, S-1
Net revenue retention (Sept. 30)About 900%S-1, CNBC
Stock-based compensation, H1$2.2 billionThe AI Insider

Two points deserve emphasis. The $80.7 billion headline loss is real in an accounting sense, but most of it comes from re-measuring convertible notes and similar instruments at fair value, which is non-cash and can swing as the company's own valuation changes. The better window into operations is the $693 million operating loss, down from $2.1 billion a year earlier. Second, the growth rate is almost without precedent for a company of this size. H1 revenue of $16.3 billion compares with $1.16 billion a year earlier, about 14 times, and second-quarter revenue was about $11.6 billion against about $4.7 billion in the first quarter.

Arithmetic on what the valuation expectations imply
Q2 2026 revenue (S-1)about $11.6 billion, or about $46 billion annualized (x4)
Multiple at $1.5 trillionabout 32x annualized Q2 revenue
Multiple at $2 trillionabout 43x annualized Q2 revenue
Q3 2026 revenue (reported by the FT via Reuters, not in the S-1)about $16.8 billion, or about $67 billion annualized
Multiple at $1.5 trillionabout 22x annualized Q3 revenue
Multiple at $2 trillionabout 30x annualized Q3 revenue
Annualizing one quarter is a rough yardstick, and the company defines its own run-rate metric differently (28 days of API usage times 13, plus monthly subscriptions times 12, per coverage of the filing). The Q3 figure comes from press reports of unfiled results. The valuations are expectations reported in the press, not filing figures, and a revenue multiple says nothing about profitability.

What worries careful readers

The bubble debate, in the speakers' own words

"Bubble" is a loaded word, so it helps to separate three different claims. One is that the technology is real and demand is real. A second is that the prices investors are paying already assume too much. A third is that the spending required to serve demand may outrun the revenue. People can accept the first and still worry about the second and third.

Who is saying what (as reported by CNBC and others, Oct. 22-27, 2026)
SpeakerPositionNotes
Denny Fish, Janus HendersonCalled it a historic moment but said details are still missing: "The devil is in the details"Portfolio manager
Erica Brescia, Redpoint VenturesSaid it is "incredible" how well Anthropic is performing so early in its lifeRedpoint is an Anthropic investor, per CNBC
Gil Luria, D.A. DavidsonIf priced appropriately, performance is likely to be excellent; called Anthropic "the premier AI company"Conditioned his view on the deal being priced appropriately
Roger McNamee, Elevation PartnersCalled the IPO "a very dangerous thing" for investors and the S-1 "a work of speculative fiction"Early Facebook and Google investor
David Trainer, New ConstructsCalled it "a massive scam for Wall Street," citing commitments and liabilities far above cash and opaque insider-sale arrangementsHis firm had earlier called it the most absurd IPO of 2026
Ebube Jones, Barchart (via Yahoo Finance)Said Anthropic is "a business worth watching, but not one worth buying at any price," citing concentration and compute commitmentsOpinion, written Oct. 22 about the draft S-1

The wording varies slightly between outlets that quoted the same interviews, so treat quotations as reported, not as transcripts. Two things stand out. Both sides cite the same numbers: growth and margin improvement for the optimists, losses, concentration, and commitments for the skeptics. And both sides make an argument about price, not about whether AI works. Even the bullish voice, Luria, conditioned his view on the deal being priced appropriately.

A real-time comparison: how SpaceX's AI spending was received

There is one live data point on how public markets treat capital-hungry AI stories. SpaceX, whose business now includes an AI compute segment, fell about 14% after its first quarterly report in August, which highlighted steep operating losses and heavy AI-infrastructure spending, per a Reuters report. It is a different company in a different business, and one reaction proves little. But it shows that public-market investors can reprice quickly when spending outruns the story. For the other side of the SpaceX experience, see our IPO best-practices article.

OpenAI is the other name in the background. Reuters reported it filed confidentially in June 2026, and it has reportedly been raising money at a valuation above $1 trillion. Sam Altman told Fortune in September that "right now would be an ill-advised moment to go public," and said an OpenAI IPO would be "not 2026."

What to watch before the deal prices

None of this tells you what the stock will do. It tells you what questions the filing raises, and which numbers to check for yourself.

Common questions

When is the Anthropic IPO?

Not confirmed. Anthropic filed its public S-1 on October 26, 2026, without a price range or share count. Bloomberg-sourced coverage says marketing could start the week of November 9, and some outlets report a debut before Thanksgiving. Those timings are unconfirmed.

What is Anthropic's ticker?

The S-1 proposes a Nasdaq listing under the ticker ANT. It is final only at pricing.

How much money does Anthropic make, and is it profitable?

Per coverage of the S-1, first-half 2026 revenue was $16.3 billion, with an operating loss of $693 million. The $80.7 billion net loss is mostly a non-cash fair-value charge on convertible instruments. The company warns it may not achieve or sustain profitability.

What valuation will Anthropic have at IPO?

Unknown. The filing gives no valuation. Press reports cite expectations from about $1.5 trillion to $2 trillion, versus $965 billion in its last private round in May 2026.

Is AI a bubble?

Analysts disagree. Optimists point to real revenue growth and customer retention. Skeptics point to customer concentration, large compute commitments, and valuation multiples. The disagreement is mostly about price and the gap between spending and revenue, not about whether the technology works.

Can retail investors buy shares in the IPO?

The SEC says most individual investors buy after trading begins, and that underwriters distribute most IPO shares to institutional and high-net-worth clients. How much retail access Anthropic's deal will include has not been detailed in the reports we reviewed. Check your broker's IPO-access program.

This article is for informational purposes only and does not constitute investment advice, and it is not a recommendation to buy, sell, or hold Anthropic (ANT), SpaceX (SPCX), or any other security. Anthropic is the maker of Claude, the AI model used to help research and draft this article; see the disclosure above. Figures come from news coverage of the S-1 and are not independently verified, and the offering terms are not final. Consult the SEC filing and a licensed financial advisor.
Sources