A reported annualized run rate tells you how fast a business is moving right now. A registration statement tells you what it actually earned. Those are different questions — and a prospective IPO filing is where the distinction stops being theoretical.
REPORTED FIGURES — NOT CONFIRMED, NOT AUDITED, NOT FROM ANY FILING
$9B
Reported annualized run rate, end of 2025
$65B
Reported annualized run rate, end of July 2026
>$100B
Run rate expected to exceed this in 2026, as reported
~$2T
Reported possible IPO valuation — a possibility, not a set term
Not investment advice. Not a recommendation. Not an offer or solicitation to buy or sell any security. No offering document is publicly available at the time of writing. All figures below are reported by news organisations — not confirmed by the company, not audited, and not drawn from any filing. Anthropic confirmed on 1 June 2026 that it had confidentially submitted a draft registration statement on Form S-1 to the SEC; a confidential draft is not a public document. Read the registration statement itself if and when it is publicly filed.
Added after publication — the one company-confirmed fact in this story. On 1 June 2026 Anthropic announced that it had confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission, saying this “gives us the option to go public after the SEC completes its review”, that any offering “will depend on market conditions and other factors”, and that “the number of shares to be offered and the price have not yet been set”. That is the company’s own statement, and it is what makes the current reporting legible: a confidential draft already under review is why a public release can be described in weeks rather than months. A confidential submission is not a public filing, no exchange was named in that announcement, and nothing in it sets a date, a price or a share count.
Reporting by The New York Times, carried by Bloomberg and Axios, states that Anthropic’s annualized revenue run rate is expected to exceed $100 billion this year. According to that same reporting, the figure stood at $9 billion at the end of 2025, rose to $65 billion by the end of July 2026, and is now on a trajectory past $100 billion. Every one of those numbers is reported by a news organisation. None is confirmed by the company, none is audited, and none is drawn from any filing.
The same reporting says a public IPO filing could be released in the coming weeks, that shares could begin trading as soon as November, that an offering could value the company at about $2 trillion, and that it could raise as much as $100 billion. The reporting itself says the timing could still change. Each of those is a reported possibility — not a set term, not a priced deal, not a filed range.
That combination — extraordinary momentum metrics arriving just ahead of a prospective disclosure regime — is the structural story here. And to read it clearly, one distinction has to come first.
The key insight: An annualized run rate is a recent period multiplied out to a yearly figure. It is not revenue recognised over a year. The gap between those two quantities is widest precisely when growth is fastest — because a fast-growing company spends most of the year at levels far below where it finishes. Stating this is not a criticism, not an accusation of exaggeration, and not a suggestion that anyone is misleading anyone. It is simply what the metric means.
The gap between a run rate and a year of revenue is widest exactly when growth is fastest. Neither number is false; they answer different questions.
The Structural Read
On the reported figures, Anthropic’s run rate moved from $9 billion at year-end 2025 to $65 billion by July 2026 — a trajectory that implies the business was operating at substantially lower levels for much of the intervening period. A company on that reported path will have recognised far less revenue for the year than the headline numbers suggest. That is not a flaw in the reporting. It is an inherent property of the metric: a run rate answers how fast is this business moving right now, not what did it earn over a defined period. Those are genuinely different questions, and using the answer to one as a proxy for the other produces a conclusion that neither figure supports.
No recognised-revenue figure is estimated, derived, or implied here. None is established by any public source. What is established is the structural gap between metric types — and why that gap will become impossible to paper over once a registration statement enters the picture.
Business Engineer — FDE Framework
Where a Momentum Metric Meets a Disclosure Regime
A registration statement reports audited recognised revenue for defined periods, prepared on a basis that a reader can verify and a regulator can examine. A run rate does not survive that translation unchanged — not because it is false, but because the disclosure regime asks a different question. The moment a company moves from private momentum reporting to public filing, the metric that defined its narrative must be restated in units the regime accepts. Neither number is wrong. They answer different questions, and a reader who treats them as equivalent will reach a conclusion that neither supports. This is the structural tension any prospective filing will have to resolve in plain arithmetic.
There is also a comparison almost every reader will attempt this week, and it is worth naming explicitly before it circulates as fact. Figures for other AI companies that have appeared recently are reported revenue figures — projected or actual — and a run rate is not the same quantity. Setting the two side by side to rank businesses produces a result that neither number supports, because both are denominated in dollars and both describe a year, yet they measure fundamentally different things. No other company’s figure is named here, no companies are ranked, and nothing here says who is ahead or behind. The general principle carries well beyond this story: when two organisations report progress in different units, the comparison that feels most obvious is almost always the one that is wrong. The burden sits with the reader to establish which unit each figure is in before doing any arithmetic at all.
Three Implications
IMPLICATION 1 — THE FILING IS THE REAL DATA EVENT
A registration statement, if and when one is publicly filed, will report audited recognised revenue for defined periods. That document — not the current wave of reported run-rate figures — is the first moment at which the scale of the business can be assessed in standardised, comparable units. The filing matters more than the report precisely because the disclosure regime does not accept momentum metrics as a substitute for recognised revenue. Readers who track AI companies closely should treat any prospective filing as the primary data event, and the current reporting as directional context rather than settled fact.
IMPLICATION 2 — FAST GROWTH WIDENS THE METRIC GAP, NOT NARROWS IT
The faster a business grows, the larger the difference between its end-of-period run rate and its full-year recognised revenue — because it spent the earlier months of the year at substantially lower levels. The reported trajectory here (from $9 billion to $65 billion in roughly seven months, then continuing higher) is precisely the kind of growth profile where the gap is widest. This is structural, not company-specific. Any business with a comparable growth curve would face the same arithmetic. Readers should hold the two measures in separate mental accounts and resist collapsing them into a single number.
IMPLICATION 3 — THE CONDITIONALS IN THE OFFERING REPORT ARE LOAD-BEARING
As reported: a filing could be publicly released in the coming weeks; shares could begin trading as soon as November; an offering couldvalue the company at about $2 trillion and could raise as much as $100 billion. The reporting itself says the timing could still change. Each “could” is doing real work. Nothing here predicts whether any offering happens, characterises demand or reception, or compares the size of any potential offering to any other. Those conditionals are not journalistic hedging — they accurately describe the state of a process that has not produced a public filing, a priced deal, or a filed range. Treat them as precise.
Business Engineer Framework
The FDE Framework — Founders, Distributors, Enablers
The FDE Framework maps where companies sit in the AI stack and which position captures durable value. Understanding how a foundation-model business like Anthropic fits — and how its reported financials should be read as it moves toward a public disclosure regime — requires knowing which layer of the stack a metric describes, not just its dollar denomination. The Map of AI traces all nine layers, including the infrastructure, model, and distribution positions where these numbers ultimately get stress-tested.
The reported run-rate figures are striking by any measure, and the reported trajectory — from $9 billion to $65 billion to a pace exceeding $100 billion, all within a single year as reported by news organisations — describes a business moving at a speed that few precedents exist to calibrate. But the structural moment here is not the number itself; it is the collision between that number and the disclosure regime a prospective public offering requires. A run rate tells you where a business is right now. A registration statement will tell you what it earned. Those are different answers to different questions, and if an IPO filing does arrive — as reported, possibly in the coming weeks, with timing that could still change — that document will be the first time the story gets told in the only unit that a regulator, an auditor, and ultimately a public market will accept.
Sources: Bloomberg / The New York Times, September 18, 2026. All figures reported by news organisations. None confirmed by the company, none audited, none drawn from any filing. This is businessanalysis of published reporting. It is not investment advice, not a recommendation, and not an offer or solicitation to buy or sell any security. No offering document is publicly available at the time of writing.
This is not investment advice, not a recommendation, and not an offer or solicitation to buy or sell any security. No offering document is publicly available at the time of writing. Every figure above is reported by a news organisation — not confirmed by the company, not audited, and not drawn from any filing. The figures described are an annualized revenue run rate, which is a recent period multiplied out. A run rate is not revenue earned over a year, and a company growing at the reported pace will have recognised materially less revenue for the year than these headline figures. That is a property of the metric, not a criticism of anyone, not an accusation of exaggeration, and not a suggestion that anyone is misleading anyone. No recognised-revenue figure is estimated, derived or implied above, and nothing above says what any filing will contain. The valuation and offering-size figures are reported possibilities rather than set terms, a priced deal or a filed range, and the reporting states the timing could still change. Nothing above predicts whether any offering occurs, characterises demand or reception, or compares any offering to another. No other company’s figure is named above, no companies are ranked, and nothing above says who is ahead or behind — a run rate and a revenue figure are different quantities and cannot be compared directly. No profit, loss, margin, cash-flow, cost, customer, contract, segment, exchange, ticker, underwriter, share-count, prior-valuation or prediction-market figure appears above.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
