The AI trade split by name on 9 November 2026: Nvidia had lost about $184 billion of market value by 10:14 a.m. Eastern while AMD, Broadcom, Alphabet and the TSMC ADR had added about $376 billion, by our arithmetic from Nasdaq quotes.
As we read it, the rest of the day’s coverage priced AI by the unit: five AI price moves that all bill by usage, Anthropic’s IPO marks, choosier lenders, power bought and refused, and banks closing accounts to outside agents. Each item links to the full piece, which names its sources.
Business Pill · CONCENTRATION
An evergreen Business Engineer explainer on concentration, when a few big pieces make the total. 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 (9 Nov 2026).
The key insight: As we read it, the day’s coverage priced AI by the unit: by 10:14 a.m. Eastern, AMD, Broadcom, Alphabet and the TSMC ADR had added about $376 billion of market value while Nvidia had lost about $184 billion, by our arithmetic, and five AI price moves on the same day all billed by usage, by our count. We do not say one explains the other.
Nvidia had lost about $184 billion of market value by 10:14 a.m. Eastern on 9 November, by our arithmetic, as Nasdaq’s quote showed its shares at $179.00, down 4.09% from Friday’s $186.64 close. In the same minute, AMD, Broadcom, Alphabet and the TSMC ADR had added about $376 billion between them, 2.04 times what Nvidia lost, by our arithmetic. Read: the five moves.
Before the open, Nvidia traded at $177.04 at 4:20 a.m. Eastern, down 5.14%, about $231 billion of market value on the 24.1 billion shares in its August 10-Q, by our arithmetic, while AMD, Broadcom and the TSMC ADR traded higher in the same quote window. Read: the pre-market fall.
Analysts split. Aletheia Capital downgraded Nvidia from Buy to Hold and withdrew its $340 price target, according to a post on X by @DeItaone, citing an estimate of $165 billion of financial commitments; Nvidia’s own 10-Q lists maximum guarantee exposure of $108.5 billion as of 26 July 2026. Bernstein said demand for Nvidia’s products seems unlikely to be affected and rates the stock “outperform” with a $400 target, Investing.com reports. Read: the Aletheia call and Bernstein’s view.
New Street Research puts up to 10% of Nvidia’s revenue as directly or indirectly exposed to China, with 6% as legal sales, CNBC reported; Nvidia’s 10-Q shows China (including Hong Kong) at $7,880 million of $96,221 million in the quarter to 26 July, 8.2% by our arithmetic. Read: the two China measures.
Elsewhere on the tape, Citi cut US equities and global tech to Neutral from Overweight, Investing.com reported; Oppenheimer upgraded AMD to Outperform with a $600 target, according to StreetInsider; and Taiwan’s Taiex closed at 41,661.70, up 4.9%, Taiwan Stock Exchange data show. Read: Citi’s call, the AMD upgrade and the Taiex.
The AI trade split by name: by 10:14 a.m. Eastern on 9 November, AMD, Broadcom, Alphabet and the TSMC ADR had added about $376 billion of market value while Nvidia had lost about $184 billion, our arithmetic from Nasdaq intraday quotes and SEC filing share counts. Intraday figures, not closing prices.Nvidia’s market value below the 6 November close at four Nasdaq quotes on 9 November: about $231 billion at 4:20 a.m. and $234 billion at 4:33 a.m. Eastern (pre-market), $204 billion at 9:53 a.m. and $184 billion at 10:14 a.m. Intraday quotes, not closing prices; values are our arithmetic on the 24.1 billion shares in Nvidia’s August 10-Q.
Priced by the Unit
Meta, Anthropic, Google, ClickUp and Apple each set or changed an AI price on 9 November, and all five bill by usage: tokens, calls, credits or a prepaid pack. By our count, two of the five add that usage on top of a seat or a subscription, and none is a new per-seat price. Read: the five pricing moves.
Meta launched Muse Ember at $0.10 per million input tokens and $0.50 per million output tokens, according to its research blog; Muse Wonder lists at $4.00 and $20.00, 40 times as much, by our arithmetic. Anthropic launched Claude Fable 5.2 and Claude Mythos 5.2 at $8 and $40, against $10 and $50 for the 5.1 models, 20% less on every token line, by our arithmetic. Read: Muse Ember and Fable 5.2.
Apple is selling a $49.99 one-time intelligence usage pack inside Logic Pro, 9to5Mac reported, about 3.8 times the $12.99 monthly Creator Studio price, by our arithmetic. ClickUp prices AI Credits at $0.001 each, so the 10,000 credits included with each $100 Max seat every month are worth $10, by our arithmetic. Read: Apple’s usage pack and ClickUp’s credits.
Strings inside a Samsung app list a Galaxy AI Premium Max plan at $4.99 a month, according to a SammyGuru teardown; Samsung has not announced it. Duolingo’s chief executive Luis von Ahn says the AI inference cost of a Video Call fell from about 30 cents at launch to well under one cent, mainly by moving to open source or open weight models. Read: Samsung’s plan in code and Duolingo’s call cost.
Anthropic’s IPO Week
Bloomberg says Anthropic can begin formally marketing its IPO as early as Tuesday, 10 November, and Bloomberg reported, as relayed in a post on X, that the company is seeking a lower-than-expected valuation target; its filings still carry no price range. Morningstar values Anthropic’s equity at $1.7 trillion in its base case, 36% above the $1.25 trillion midpoint of the range potential IPO buyers were discussing, as Bloomberg reported on 6 November, by our arithmetic. Read: the roadshow timing and Morningstar’s $1.7 trillion.
By our count, nine sponsors filed proposals for Anthropic funds before the first IPO filing EDGAR lists for Anthropic, the earliest 112 days ahead of its 1 June draft. Anthropic’s research and development spending was 4.0 times its cost of revenue in the first nine months of 2026, up from 2.0 times a year earlier, by our arithmetic from its amended S-1. Read: the ETF filings and the R&D ratio.
Claude rose six places to No. 25 on Similarweb’s list of the most-visited websites in October 2026, the firm said on X, while Perplexity fell 26 places to No. 392. Read: the October ranks.
Credit Gets Choosier
Some banks that were lining up to fund AI deals a few weeks ago are now reluctant to finance additional deals, Reuters reported, citing banking and credit sources, even when deals are backstopped by names such as Anthropic and Google. Bank of America’s Yuri Seliger puts the widening in 10-year hyperscaler spreads at about 50 basis points this year against roughly 20 for non-AI industrials. Read: the lenders’ pullback.
Morgan Stanley is keeping its hyperscaler capex forecasts at $1.6 trillion in 2027 and $1.8 trillion in 2028, according to its note as shown in images posted on X, with Google the largest spender at $375 billion in 2027; the note shifts 5 gigawatts of US data center projects from 2028 to 2029. Read: the $1.6 trillion call.
Bank of England Governor Andrew Bailey said in a speech published on 9 November: “Bubble spotting is an imprecise science. But, let’s recognise that such a risk exists, which we should do”. The Bank’s Financial Policy Committee record cites a JP Morgan estimate of about $4.1 trillion of debt-financed AI capital expenditure between 2026 and 2030, about $820 billion a year by our arithmetic. Read: Bailey’s speech.
By our classification, three of four banks quoted after the US midterms point to a slower pace or delays for the AI data center buildout. Alphabet has had talks about Hong Kong and Singapore dollar bonds worth $2.5 billion to $3 billion in total, by our arithmetic from Bloomberg’s report, and SoftBank Corp.’s head is preparing for “a possible slowdown in AI data center demand in the US”, Bloomberg reported. Read: the bank scorecard, Alphabet’s Asia bonds and SoftBank Corp..
CoreWeave signed a multi-year agreement with Wayve a day before it reports third-quarter results on Tuesday, 10 November; its latest disclosed revenue backlog was $103.7 billion at 30 June. Read: the Wayve deal.
Power, Bought and Refused
By our count, seven actions pushing back against data centres were dated or reported between 31 October and 9 November in five countries, four of them on 9 November. Among them, Statnett, Norway’s grid operator, paused further grid reservations for data centres: it says 800 MW are connected, while connected, reserved, queued and requested capacity adds up to 32,100 MW, by our arithmetic, against Norway’s maximum power consumption of slightly over 25,000 MW. Read: the pushback count and Norway’s pause.
Power companies reported deals on the same day. Fervo says a 396-megawatt agreement with Google lifted its contracted capacity 60% in a quarter to 1,054 megawatts, by our arithmetic; Constellation says its new Google and Amazon agreements add more than 1,000 megawatts; and NRG agreed to sell 4.9 GW of gas plants for $4 billion, about $816 per kilowatt by our arithmetic. Read: Fervo, Constellation and NRG.
In the United States, delivery of a new gas turbine now takes six years on average, a Wood Mackenzie analyst told Reuters; a refurbished unit costs up to 48% less than a new one, by our arithmetic from the prices Reuters cites. Read: the turbine market.
Banks Close the Door to Agents
HSBC, Nationwide, NatWest, Barclays and Lloyds have moved to block AI agents from customer accounts, The Telegraph reported. By our count, 15 companies and banks named in our coverage now restrict outside AI agents, 12 financial firms and three retailers, and 13 of the 15 were reported between 27 October and 8 November. Read: the UK banks and the count of 15.
Marks on Anthropic’s equity value: Morningstar’s bull case above $3 trillion, base case $1.7 trillion and bear case $500 billion (9 November), against the $1.25 trillion midpoint of the range IPO buyers were discussing, as Bloomberg reported on 6 November; the 36% gap is our arithmetic.
The Structural Read
As we read it, the selling on the tape stayed in one name: the four AI chip names we tracked added 2.04 times what Nvidia lost by 10:14 a.m. Eastern, by our arithmetic, in intraday quotes rather than closing prices.
As we read it, sellers of AI priced by the unit: Meta and Anthropic by the token, Google by the call, ClickUp by AI credit and Apple by a $49.99 usage pack, with none adding a new per-seat price, by our count.
As we read it, the money behind the build got more selective at the same time: Reuters reported some lenders reluctant to fund more AI deals, Bailey said a bubble risk exists, and Norway’s grid operator paused new data centre reservations. We do not say why prices moved.
Andrew Bailey, Governor of the Bank of England, in a speech published by the Bank on 9 November 2026, as quoted in our piece
“Bubble spotting is an imprecise science. But, let’s recognise that such a risk exists, which we should do”
Three Implications
AI CHIP INVESTORS By 10:14 a.m. Eastern, AMD, Broadcom, Alphabet and the TSMC ADR had added about $376 billion while Nvidia had lost about $184 billion, by our arithmetic.
AI BUYERS Anthropic’s Fable 5.2 lists at $8 and $40 per million tokens, 20% less than 5.1, and Meta’s Muse Ember at $0.10 and $0.50, according to their own pages.
ANTHROPIC IPO BUYERS Morningstar’s $1.7 trillion base case sits 36% above the $1.25 trillion midpoint of the range IPO buyers were discussing, by our arithmetic; Anthropic has not filed a price range.
The analysis argues that “The cost of intelligence is collapsing while the cost of deploying intelligence is exploding”, and that the token is “the unit of price the lab charges”.
As we read it, the day’s price moves sat on that token unit: per-token list prices from Meta and Anthropic, and credits and usage packs layered on top of seats and subscriptions at ClickUp and Apple. We do not say one caused the other. The analysis was published in June 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 a split by name and by unit, are our grouping, not a finding.
We do not claim that any one story moved prices or caused another. The market-value figures are intraday and pre-market quotes, not closing prices; several items rest on outlet reports citing people familiar with the matter, on posts on X or on app code; and Anthropic has not filed an IPO price range.
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 pricing AI this week, two numbers sit side by side: about $376 billion added by four AI chip names against about $184 billion lost by Nvidia by 10:14 a.m. Eastern, by our arithmetic, and five AI price moves on one day, all billed by usage, by our count.
For the frame behind the pricing thread, read The CFO’s Guide to the Token Economy on The Business Engineer, on why cheaper tokens can still mean bigger AI bills: 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 8 and 9 November 2026, linked above; every story was built from its own primary source, which it cites and links. We re-read each piece on 9 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.
