Clarification (10 Nov, 11:3x UTC): which share count this piece uses. Anthropic’s S-1/A, Amendment No. 2 repeats the sentence this piece is built on: the 846,840,960 Public Benefit Fund shares are “5% of our fully diluted shares as of June 30, 2026”. So the 16.94 billion base below is a fully diluted count before the IPO, not shares outstanding. The filing says 13,086,672,593 shares will be outstanding after the IPO and placement. The filed range is $70 to $78, not the retracted $70 to $80. On this piece’s method, the filed range gives $1.27 trillion to $1.41 trillion; on the filed outstanding count it is $917 billion to $1.02 trillion, all by our arithmetic. Each $10 of IPO price is about $169 billion on the fully diluted base and about $131 billion on the outstanding count. Details in our piece on the filed terms.
Anthropic’s own S-1/A puts its fully diluted share count at about 16.94 billion as of 30 June, by our arithmetic. On that base, the $70 to $80 price range Anthropic sent bankers and then retracted works out to $1.19 trillion to $1.35 trillion. Add the new shares and the placement shares in that retracted filing copy, as Reuters reported, and it comes to $1.27 trillion to $1.45 trillion.
That is the same span a widely shared post by Wall St Engine gave at 01:38 UTC on 10 November, which said it was a “Calculation by Claude”. In our piece on the retracted range we wrote that the share count behind that figure could not be checked against the public filing. The S-1/A’s own numbers allow that check, with caveats set out below.
Business Pill · Run-Rate Revenue
A one-minute explainer of run-rate revenue, a measure often used to describe fast-growing AI companies ahead of listings. It teaches the general idea only and does not discuss Anthropic’s share count.
The key insight: Anthropic’s S-1/A implies about 16.94 billion fully diluted shares at 30 June, by our arithmetic. On that base the $70 to $80 range Anthropic sent bankers and then retracted, per Reuters, is $1.19 trillion to $1.35 trillion, and $1.27 trillion to $1.45 trillion with the new and placement shares in the retracted copy, matching the relayed estimate.
The S-1/A filed on 2 November leaves the price range and the shares outstanding after the offering blank. It does give one anchor. Anthropic says it “contributed 846,840,960 shares of our Class N common stock, representing 5% of our fully diluted shares as of June 30, 2026 (assuming full conversion of the Amazon Notes, as defined below)” to the Anthropic Public Benefit Fund.
Anthropic’s implied value at $70 and $80 a share, the range in the filing copy Anthropic later said had inaccurate pricing details (Reuters), on the 16.94 billion fully diluted shares implied by its S-1/A at 30 June; then adding the about 843.3 million new shares and the $26 billion placement shares in that copy; against the estimate posted by Wall St Engine. Data: Anthropic S-1/A, 2 Nov 2026; Reuters, 10 Nov 2026; Wall St Engine on X; our arithmetic.
The Retracted Range on That Base
Reuters’ updated report says the filing copy Anthropic later said had inaccurate pricing details showed 866.7 million Class A shares: about 843.3 million sold by Anthropic and 23.3 million by existing shareholders. It also showed Google, Nvidia, AMD, Broadcom, Micron, Microsoft and TSMC buying $26 billion of Class N common stock at the IPO price, in a private placement closing right after the IPO, which our sibling piece covers.
By our arithmetic, in three steps: the 16.94 billion base at $70 and $80 is $1.19 trillion to $1.35 trillion. Adding the 843.3 million new Anthropic shares gives 17.78 billion shares and $1.24 trillion to $1.42 trillion. The $26 billion placement buys about 371 million shares at $70 and 325 million at $80; adding those gives $1.27 trillion to $1.45 trillion. The 23.3 million shares sold by existing holders are already in the base, so they add nothing.
Wall St Engine’s $1.27 trillion and $1.45 trillion imply about 18.14 billion and 18.13 billion shares, by our arithmetic, roughly 1.2 billion more than the 30 June base. The third step lands within about 0.1% of both figures. The post does not say how it was calculated, so this shows the figures can be rebuilt from filed and reported numbers, not that the poster used this method.
Against the Other Marks
The S-1/A says Series H-1 preferred was sold at $589.0095 per share, and that the conversion ratio for its preferred stock was “10 : 1” at 30 September except for three Series D series. At that ratio, Series H-1 is about $58.90 per common share, so $70 is 18.8% above it and $80 is 35.8% above, by our arithmetic.
The $1.19 trillion low end is 22.9% above the $965 billion Series H post-money valuation in our funding history. Morningstar’s $1.7 trillion base case, in our Morningstar piece, is above every figure here: the $1.45 trillion top is 14.7% below it, by our arithmetic.
The Structural Read
As we read it, the relayed $1.27 trillion to $1.45 trillion can be rebuilt from filed and reported numbers: the S-1/A share anchor plus the new and placement shares in the retracted copy lands within about 0.1% of it, by our arithmetic.
As we read the S-1/A, each $10 on the IPO price moves the implied value by about $169 billion on the 30 June base, by our arithmetic.
Anthropic, PBC Form S-1/A, 2 November 2026
“contributed 846,840,960 shares of our Class N common stock, representing 5% of our fully diluted shares as of June 30, 2026 (assuming full conversion of the Amazon Notes, as defined below)”
Three Implications
FOR IPO BUYERS The range was retracted and no new S-1/A was on EDGAR at 07:16 UTC on 10 November; on the 30 June base, $70 and $80 are $1.19 trillion and $1.35 trillion, by our arithmetic.
FOR THE PLACEMENT The $26 billion Class N placement at the IPO price would add about 325 million to 371 million shares at $80 to $70, by our arithmetic, on the retracted terms Reuters reported.
FOR THE VALUATION MARKS The $1.19 trillion low end is 22.9% above the $965 billion Series H post-money; the $1.45 trillion top is 14.7% below Morningstar’s $1.7 trillion base case, by our arithmetic.
The framework argues: “That is the structural read, and it changes the valuation logic at every layer beneath it.”
As we read it, Anthropic’s own cap-table anchor turns a relayed trillion-dollar estimate into arithmetic anyone can check, and the open question becomes the share count in the next filing. The framework predates the retracted range and does not discuss it.
Update (10 Nov): a columnist’s opinion on the price. In a Bloomberg Opinion column (syndicated text), columnist Chris Hughes argues there “should be no shame in an IPO valuation closer to $1 trillion than the $2 trillion originally envisaged.” His method: value Anthropic’s prospective 2029 revenue at around 7.5 times, discount that back at 15% a year to about $1.5 trillion today, then apply a 20% to 30% IPO discount, which lands on $1.1 trillion to $1.2 trillion.
He adds that haggling for the best price would put post-float performance at risk. This is opinion and his arithmetic, not a reported price range; the share-count mark above is unchanged by it.
What Is Not Established
The base is a 30 June figure that assumes full conversion of the Amazon Notes. The S-1/A does not break that figure down; it lists 1,774,342,693 options outstanding at 30 September with a weighted-average exercise price of $2.91 per share, and how many of them the base counts is not stated. The base does not capture grants, conversions or conversion-ratio changes after 30 June. The S-1/A also says the Public Benefit Fund’s shares are not considered outstanding for accounting purposes while the fund holds them.
The $70 to $80 range was retracted: Anthropic told bankers the pricing details were inaccurate and that it would update the filing, Reuters reported, and Anthropic declined to comment. No new S-1/A was on EDGAR when we checked at 07:16 UTC on 10 November, so the new range, the final share count and the shares outstanding after the offering are not public. The timetable is in our piece on the planned trading date.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
On the 16.94 billion fully diluted shares implied by Anthropic’s own S-1/A, the retracted $70 to $80 range is $1.19 trillion to $1.35 trillion, by our arithmetic. Adding the new shares and placement shares in the retracted copy takes it to $1.27 trillion to $1.45 trillion. That matches the relayed estimate. Each $10 of IPO price is worth about $169 billion on the 30 June base.
A note on sourcing. The share anchor, the Series H-1 price, the conversion ratio and the option count are from Anthropic’s S-1/A of 2 November 2026 on SEC EDGAR. The retracted terms are from Reuters’ updated report of 10 November, read in full via CNA syndication; Anthropic declined to comment to Reuters. The relayed estimate is quoted in full from the X post. The fully diluted base, the implied valuations, the implied share counts and the percentages are our arithmetic. Anthropic was not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
