The valuation that investors discussed for Anthropic’s initial public offering fell from $1.5 trillion to $1.6 trillion to $1.2 trillion to $1.3 trillion in four days. At the midpoints, $1.55 trillion against $1.25 trillion, that is $300 billion, or 19.4%, lower, by our arithmetic.
In the same week, sellers of scarce AI inputs reported strength: Counterpoint Research said global DRAM revenue grew 369% year on year in the third quarter, to nearly $200 billion, while bit shipments rose only about 4%.
Business Pill · Concentration
A short explainer on concentration. It teaches the general idea only and does not discuss these companies.
The key insight: As we read the figures, the valuation investors discussed for Anthropic fell $300 billion at the midpoints in four days, while Counterpoint put third-quarter DRAM revenue up 369% on about 4% more bits, by our arithmetic.
The Wall Street Journal reported on Monday, 2 November, that institutional investors had told Anthropic and its advisers that its valuation should be “around $1.5 trillion to $1.6 trillion”, as we covered at the time.
A Bloomberg headline published at about 13:14 UTC on 6 November said investors were discussing $1.2 trillion to $1.3 trillion. Bloomberg’s earlier article that day put investors’ fair-value estimate at “around $1.4 trillion to $1.5 trillion”; measured from that midpoint, the drop is $200 billion, or 13.8%, by our arithmetic.
No price range has been set. These are figures investors discussed, as reported, not a price.
The three valuation ranges investors discussed for Anthropic’s IPO, as reported by The Wall Street Journal on 2 November and Bloomberg on 6 November 2026. Midpoints $1.55 trillion to $1.25 trillion = $300 billion, or 19.4%, lower: our arithmetic.
Who Sold Scarce Inputs
Counterpoint said prices drove the DRAM revenue growth: average selling prices reached $1.9 per gigabit, and Samsung kept a 38% share. These are third-quarter figures, published on 5 November.
GE Vernova’s latest quarterly filing shows customers ordered 57 aeroderivative gas turbines, which are built from jet engines, in the three months to 30 September, up from 9 a year earlier, while it sold 13.
Missouri utility Ameren said on its third-quarter call that it plans “$38.1 billion in capital expenditures from 2027 through 2031”, tied to “2.8 gigawatts of signed electric service agreements”, as we reported.
Bond buyers also lined up for a hyperscaler. Meta’s first euro bond drew €53 billion of orders for a sale of up to €9 billion, according to the markets feed First Squawk, about 5.9 times the maximum size, against about 3.9 times for its $15 billion dollar sale a day earlier, by our arithmetic.
Who Borrowed
Six listed AI cloud operators lost an average of 20.9% from their 30 October close to their 5 November close, against 3.0% for the S&P 500, by our arithmetic on Nasdaq official closes.
Bloomberg reported that SoftBank Group’s dollar bond due in 2032 fell 0.9 cents, its biggest drop in 10 weeks, and that its shares fell as much as 16%. Of 14 SoftBank dollar note lines that last traded on 4 or 5 November on FINRA’s TRACE system, 13 printed below par, our count shows.
The split was not clean. In the first US session after the midterms, on 4 November, ten data-center builders and power suppliers were down an average of 11.8% at 11:51 a.m. ET, by our arithmetic on Nasdaq.com intraday prices, while four hyperscalers averaged a 5.1% gain. Those are intraday prices, not closes.
As we read the two Bloomberg reports of 6 November, most of the drop came on one day: $200 billion of the $300 billion, measured from the $1.45 trillion midpoint, by our arithmetic.
As we read the input figures, each is a count of orders or revenue already booked: 57 aeroderivative turbines ordered against 9, 2.8 gigawatts of signed agreements, and a euro bond book about 5.9 times its maximum size.
As we read the credit figures, the borrowers named here sit between the model companies and the suppliers: AI cloud operators, SoftBank and SpaceX.
Ameren, third-quarter call
“2.8 gigawatts of signed electric service agreements”
Three Implications
FOR THE MODEL LAYER The valuation investors discussed for Anthropic fell from $1.5 trillion to $1.6 trillion to $1.2 trillion to $1.3 trillion between 2 and 6 November, as reported.
FOR THE INPUT SELLERS Counterpoint said DRAM revenue rose 369% in the third quarter on about 4% more bits, and GE Vernova reported 57 aeroderivative orders against 9 a year earlier.
FOR THE BORROWERS Six AI cloud operators fell 20.9% on average and SpaceX credit default swaps reached a record 286.56 basis points, by our arithmetic and as reported.
The analysis says: “AI exhibits genuine bubble characteristics — excessive hype, overvaluation, capex commitments that exceed the unit economics of any single application — at the same time as it exhibits genuine supercycle characteristics — real productivity gains, fundamental restructuring of profit pools, civilizational impact.”
As we read it, this week put both sides of that tension in one set of figures: a lower mark at the model layer and booked revenue at the input layer. The analysis does not discuss these companies.
What Is Not Established
The Anthropic figures are valuations investors discussed, reported by the Journal and Bloomberg; no price has been set. The DRAM, turbine and utility figures cover the third quarter and earlier, not the week of the rout. We make no claim that money moved from one group to the other; the piece sets reported figures side by side.
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.
Between 2 and 6 November, the valuation investors discussed for Anthropic’s IPO fell from $1.5 trillion to $1.6 trillion to $1.2 trillion to $1.3 trillion, $300 billion lower at the midpoints, by our arithmetic. Counterpoint put third-quarter DRAM revenue up 369% on about 4% more bits, while six AI cloud operators fell 20.9% and SpaceX credit protection hit a record.
A note on sourcing. The Anthropic ranges are as reported by The Wall Street Journal and Bloomberg. Each other figure comes from our linked piece, which cites the named primary source. Midpoints, differences and ratios are our arithmetic.
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.
