Anthropic put a price on itself on 10 November 2026: $70 to $78 a share, $968.4 billion at the midpoint on the share count it filed, by our arithmetic, with Nvidia and Google buying $10.0 billion each of a $26.0 billion placement. By 10:31 a.m. Eastern the same day, Nvidia was about $368 billion below its Friday close, by our arithmetic from Nasdaq quotes.
As we read it, the rest of the day’s coverage followed the money and the prices around that filing: Anthropic’s office tools against Microsoft, Nvidia’s tape and board, model prices with an end date, AI in the inflation data, SoftBank’s quarter, and agents opening shop. Each item links to the full piece, which names its sources.
Business Pill · IPO FLOAT
An evergreen Business Engineer explainer on IPO float, the share of a company sold in its listing. It teaches the general idea only and does not discuss the companies in this roundup.
AI Daily Briefing · video
Today’s stories in about a minute (10 Nov 2026).
The key insight: As we read it, the day’s coverage followed money around one filing: Anthropic filed a $70 to $78 range, $968.4 billion at the midpoint on its filed share count, with Nvidia and Google buying $10.0 billion each of a $26.0 billion placement, while Nvidia was about $368 billion below its Friday close by 10:31 a.m. Eastern, by our arithmetic. We do not say one explains the other.
Anthropic filed its price range on 10 November: 825 million Class A shares at $70 to $78, in an amended S-1 accepted by the SEC at 11:29 UTC. On the 13,086,672,593 shares the filing says will be outstanding after the IPO and a $26 billion placement, the $74 midpoint values Anthropic at $968.4 billion, by our arithmetic, $917 billion at $70 and $1.02 trillion at $78; on the fully diluted base our earlier pieces used, it is $1.27 trillion to $1.41 trillion. Read: the filed range.
The same filing gives the split of the $26.0 billion placement for the first time: Nvidia and Google $10.0 billion each, AMD $2.5 billion, Broadcom, Micron and Microsoft $1.0 billion each, and TSMC $500 million, all in non-voting Class N stock at the IPO price. Google, Microsoft and Broadcom put in $12.0 billion between them; the filing names $329.7 billion that Anthropic has committed to spend with those three, 27.5 times as much, by our arithmetic. Read: the placement.
Anthropic’s run-rate revenue was approximately $81 billion on 31 October 2026, up $10 billion from 30 September, the filing says; the $10 billion added in October alone is about 2.2 times its 2025 revenue of $4,585 million, by our arithmetic, though the two are not the same measure. Anthropic’s IPO book is “already covered at the $70 to $78 per share range on an allocable basis”, CNBC reported, citing people familiar with the details. Read: the run-rate and the book.
The filing also adds costs and dates. It lists about $17.4 billion of minimum payments under short-term compute contracts at 30 September and about $6.6 billion signed after that date, $24.0 billion, 4.8 times the third quarter’s $5.0 billion short-term compute bill, by our arithmetic. Expected IPO expenses are $67,684,106, not counting the underwriting discount, and legal fees and expenses of $41,479,483 are 61% of that, by our arithmetic. Read: the short-term contracts and the cost table.
On 9 November Anthropic’s board moved the service-vesting date for about 67 million restricted stock units from 20 November to 19 November, the filing says, adding that the extra stock-based compensation expense will be “material”. Up to 316 million Class A shares can be sold on the first trading day after the prospectus date, and up to 1,612 million by the 76th day, 1.95 times the shares offered, by our arithmetic. Read: the RSU change and the lock-up calendar.
Anthropic’s $26.0 billion placement at the IPO price, by buyer, from its S-1/A of 10 November: Nvidia and Google $10.0 billion each, AMD $2.5 billion, and Broadcom, Micron, Microsoft ($1.0 billion each) and TSMC ($500 million), $3.5 billion together. The sums are our arithmetic.
Into Microsoft’s Office
Anthropic on 10 November launched Claude Sheets and Claude Whiteboards in beta and made Claude Data generally available, on all paid plans. A Claude Team standard seat costs $20 a month billed annually, $3.50 below the $23.50 Microsoft lists for Microsoft 365 Business Standard with Copilot, by our arithmetic from the two price pages. Read: the office tools.
Microsoft was worth about $13 billion more than Nvidia at 12:09 p.m. Eastern on 10 November, by our arithmetic from Nasdaq quotes and SEC filing share counts; Nvidia was back ahead by 12:22 p.m. At the 6 October close, Nvidia had led Microsoft by about $1.84 trillion. Read: the brief crossover.
IBM and Microsoft announced Imagine Enterprise, a platform to “build, run and govern AI agents”, with IBM Consulting bringing “more than 6,000 prebuilt AI agents and assets”. IBM’s third-quarter 10-Q shows Consulting signings down 3.7% to $5,030 million, 0.94 times the quarter’s $5,335 million of Consulting revenue, by our arithmetic. Read: the IBM and Microsoft platform.
Nvidia: The Tape and the Board
Nvidia had lost about $368 billion of market value since Friday’s close by 10:31 a.m. Eastern on 10 November, by our arithmetic, as Nasdaq’s quotes showed its shares at $171.36, down 8.19% from the $186.64 close on 6 November. Over the same two sessions, AMD and Broadcom had added about $317 billion between them, 86% of what Nvidia lost, by our arithmetic. Read: the two sessions.
Bank of America removed Nvidia, Micron and Lam Research from its US 1 list while all three keep Buy ratings, and Wells Fargo cut its Nvidia price target to $300 from $315, Investing.com reported. FMR LLC, the parent of Fidelity Investments, reported holding 932,807,820 Nvidia shares at 30 September, down 9.09% from 30 June, in a Form 13F, by our arithmetic. Read: the bank calls and Fidelity’s filing.
Nvidia’s board, which said on 8 November it is opening an independent investigation, has no chair: its 2026 proxy statement describes “an independent Lead Director, rather than a chairperson”. By our arithmetic from the proxy, 4 of its 10 nominees have served since the 1990s, and the median nominee has served 15.5 years, against 13.5 at Apple and 10 at Amazon. Read: the board.
On policy, The Information reported that White House feedback sent to congressional staff in late October called for an AI Overwatch Act provision ordering the Commerce Department to close the remote access loophole on the chips the bill covers; by our reading, the House text introduced in December mentions remote access only twice. Read: the remote access report.
Model Prices With a Date on Them
Google made Gemini 4 Carbon available at $2 per million input tokens and $10 per million output tokens “through November 30, 2026”, then $4 and $20 from 1 December, according to its pricing page. At the introductory price Carbon costs half of Claude Opus 5.5’s list price, and from 1 December it matches Opus 5.5 per token, by our arithmetic. Read: Carbon’s price.
Perplexity says Carbon scored 0.857 on its WANDR research benchmark at $5.86 per task, “matching Claude Fable 5.2 at 40% of the cost”; a day earlier it posted 0.851 at $14.51 per task for Fable 5.2. That is 40.4% of Fable 5.2’s cost per task, by our arithmetic. Read: the WANDR scores.
AI in the Price Data
The US Consumer Price Index rose 0.4% in October, and the index for electricity rose 0.9%, 2.25 times the all-items rise, by our arithmetic from the Bureau of Labor Statistics release of 10 November. The Energy Information Administration expects commercial power sales, the sector its October outlook described as the one “which includes data centers”, to take 97 of the 151 billion kWh rise in US sales from 2025 to 2027, 64%, by our arithmetic. Read: October’s CPI and the EIA outlook.
Dallas Fed President Lorie Logan said the FOMC “needs to implement at least two more 25-basis-point rate increases”. In Sydney, the Reserve Bank of Australia’s Sarah Hunter said there is “limited evidence that AI has materially lifted aggregate productivity”, while an RBA staff note puts a further A$115 billion in real terms into Australian data centres between mid-2026 and the end of 2028. Read: Logan’s speech and the RBA.
Also on 10 November, Federal Reserve Vice Chair for Supervision Michelle Bowman said “Tariffs, energy, and AI-related price pressures each accounted for roughly one-fourth of above-target core PCE inflation over the 12 months ending in September”, and that core PCE inflation “likely moved up further to 3.3 percent in October”. Read: Bowman’s speech.
In Washington, the Senate voted 73-25 on 9 November to open debate on the Bipartisan American Affordability and Jobs Act, the permitting bill that would make large data centers pay more of the grid costs they cause, according to the Senate roll call and Reuters. Read: the Senate vote.
SoftBank’s Quarter
SoftBank Group posted a net loss attributable to owners of the parent of ยฅ1,115.3 billion for July to September, after earning ยฅ347.3 billion in April to June, according to its data sheet. The financial report books a ยฅ597.2 billion ($3.73 billion) loss after a subsidiary realigned its purchases of power-generation equipment for US AI data centers. Read: the loss.
SoftBank carried its OpenAI stake at a fair value of $99.6 billion on 30 September against a cumulative investment of $54.6 billion, 1.82 times cost, by our arithmetic. Its pro forma net asset value was ยฅ44.7 trillion at 9 November prices, 38% below the ยฅ72.3 trillion it reported for 30 June, by our arithmetic from its earnings presentation. Read: the OpenAI stake and the NAV.
Agents Get a Storefront
OpenAI launched “Businesses” in beta within ChatGPT, letting companies claim a profile and, for eligible US e-commerce and online service businesses, publish a Business Agent, Reuters reported; Digiday reported the hub and agents are free for now, 55 days after OpenAI began testing a paid version, by our arithmetic. Founder Francis Davidson wrote that OpenAI has acquired Odessia, an AI travel agent; no price was disclosed. Read: ChatGPT’s Business Agents and Odessia.
DoorDash’s own Agentic Messaging Index puts Detroit at 178 and Chula Vista at 66 for September 2026, 2.7 times apart, by our arithmetic. Going the other way, The Verge reported that eBay appears to have blocked AI agents including Meta’s Muse, and eBay’s robots file names 15 AI agents, by our count, up from 10 in a 15 May copy. Read: DoorDash’s index and eBay’s block.
Anthropic’s filed range of $70 to $78 a share on three share counts from its S-1/A of 10 November: $917 billion to $1.02 trillion on shares outstanding after the IPO, $1.12 trillion to $1.24 trillion counting options and unvested awards, and $1.27 trillion to $1.41 trillion fully diluted. Valuations are our arithmetic; Anthropic has filed a range, not a price.
The Structural Read
As we read it, the placement ties Anthropic’s suppliers to its listing: Google, Microsoft and Broadcom put in $12.0 billion, while the same filing names $329.7 billion Anthropic has committed to spend with those three, 27.5 times as much, by our arithmetic.
As we read it, the filing’s numbers point both ways: run-rate revenue of approximately $81 billion on 31 October, and $24.0 billion of minimum payments under short-term compute contracts, by our arithmetic, compute the filing says currently costs significantly more than its multi-year compute.
As we read it, AI showed up in the price data as well as in share prices: the US electricity index rose 0.9% in October, 2.25 times the all-items rise, by our arithmetic, and a Federal Reserve president called for at least two more rate increases. We do not say why prices moved.
Lorie Logan, President of the Federal Reserve Bank of Dallas, in a speech in Houston on 10 November 2026, as quoted in our piece
“It remains clear to me, given the current outlook, that the FOMC needs to implement at least two more 25-basis-point rate increases.”
Three Implications
ANTHROPIC IPO BUYERS The filed range gives $917 billion to $1.02 trillion on shares outstanding after the IPO and $1.27 trillion to $1.41 trillion fully diluted, by our arithmetic; Anthropic has filed a range, not a price.
OFFICE SOFTWARE BUYERS A Claude Team standard seat lists at $20 a month billed annually, $3.50 below Microsoft 365 Business Standard with Copilot at $23.50, by our arithmetic from the two price pages.
AI MODEL BUYERS Gemini 4 Carbon lists at $2 and $10 per million tokens through 30 November and $4 and $20 from 1 December, according to Google’s pricing page.
The analysis argues that “The $1T capex cycle is being underwritten by a $1T concentrated demand commitment from two counterparties whose combined free cash flow is deeply negative”, naming OpenAI and Anthropic.
As we read it, the day’s filing shows that link running both ways: Anthropic names $329.7 billion of commitments to Google, Microsoft and Broadcom, and those three buy $12.0 billion of its stock at the IPO price. We do not say one caused the other. The analysis was published in May 2026 and does not discuss these events.
What Is Not Established
This roundup adds no new reporting and selects from the day’s coverage. Every figure above is taken from the linked FourWeekMBA piece, each built from its own primary source. The threads, and the reading of the day as following the money and the prices around one filing, are our grouping, not a finding.
We do not claim that any one story moved prices or caused another. Anthropic has filed a range, not a price; the market-value figures are intraday quotes, not closing prices; and several items rest on outlet reports citing people familiar with the matter or on posts on X.
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.
For anyone following AI money this week, two numbers sit side by side: $968.4 billion at the midpoint of Anthropic’s filed range, by our arithmetic, with $12.0 billion of its placement coming from Google, Microsoft and Broadcom, and about $368 billion of Nvidia market value below Friday’s close by 10:31 a.m. Eastern, by our arithmetic.
For the frame behind the placement thread, read The AI Capex Map and the State of AI Hyperscalers on The Business Engineer, on how a few counterparties underwrite the build: read the analysis.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. This roundup draws on FourWeekMBA’s news coverage of 9 and 10 November 2026, linked above; every story was built from its own primary source, which it cites and links. We re-read each piece on 10 November 2026 and took every figure here from it. The threads and the figures in our charts are our own grouping and selection. Nothing here is a forecast, and nothing here is financial or investment advice.
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.
