Bloomberg reported on 4 November that shares of US tech giants jumped after the midterms, as investors bet that a Democrat-controlled Congress will slow AI capital spending, according to Techmeme’s summary of the article, which we have not read in full.
At 1:09 p.m. ET (18:09 UTC) on 4 November, the first US trading session after the midterm elections, Amazon, Microsoft, Alphabet and Meta had added about $653 billion of stock market value between them since the 3 November close, by our arithmetic from Nasdaq.com prices and the companies’ share counts. Five chipmakers, Nvidia, Micron, Broadcom, Intel and AMD, had lost about $149 billion net.
The hyperscalers’ gain was 4.4 times the chipmakers’ loss, by our arithmetic. These are intraday prices, not closing prices.
Business Pill · Capex
An evergreen Business Engineer explainer on capital expenditure. It teaches the general idea only and does not discuss the elections or any company.
Change in market value from the 3 November close to 1:09 p.m. ET on 4 November 2026, by company. Prices: Nasdaq.com; share counts: 10-Q/10-K cover pages; dollar changes are our arithmetic. Chart: FourWeekMBA.
The key insight: At 1:09 p.m. ET on 4 November, Amazon, Microsoft, Alphabet and Meta had added about $653 billion of market value since the 3 November close, 4.4 times the roughly $149 billion net that Nvidia, Micron, Broadcom, Intel and AMD had lost, by our arithmetic from Nasdaq.com prices.
Update, 5 Nov, 07:34Z: Micron at the close. The figures below this box are intraday prices from 1:09 p.m. ET. At the 4 November close, Micron finished at $795.23, down $59.87 or 7.00%, on about 35.9 million shares, according to Nasdaq.com.
At 1,131,423,212 shares outstanding on the cover of Micron’s 10-K filed 9 October, that is about $67.7 billion of market value lost on the day, by our arithmetic, against the roughly $53.5 billion we estimated at 1:09 p.m. Micron’s market value at the close was about $900 billion, by our arithmetic.
Sandisk, not one of the five chipmakers in this piece, closed at $1,037.40, down 7.91%, according to Nasdaq.com. We have not recomputed the piece’s other closing figures; the $149 billion and 4.4x numbers remain the 1:09 p.m. ET estimates.
According to Nasdaq.com, Amazon was up 7.63%, Microsoft 6.41%, Alphabet’s Class A shares 4.37% and Meta 2.99%. Micron was down 5.53%, Intel 3.17%, Nvidia 1.43% and Broadcom 1.11%, while AMD was up 0.40%.
In dollars, by our arithmetic, Microsoft added about $248.0 billion, Amazon $206.9 billion, Alphabet $150.5 billion and Meta $47.5 billion. Nvidia lost about $71.0 billion, Micron $53.5 billion, Broadcom $17.7 billion and Intel $10.2 billion; AMD added about $3.2 billion.
Microsoft’s gain alone, about $248.0 billion, was larger than the five chipmakers’ combined net loss, by our arithmetic.
Earlier in the session, at 11:51 a.m. ET, the four hyperscalers averaged a 5.1% gain while ten data-center builders and power suppliers were down an average of 11.8%, by our arithmetic from Nasdaq.com prices.
Change in stock market value from the 3 November close to 1:09 p.m. ET on 4 November 2026: Amazon, Microsoft, Alphabet and Meta about +$653 billion combined; Nvidia, Micron, Broadcom, Intel and AMD about -$149 billion net. Prices from Nasdaq.com, share counts from 10-Q/10-K cover pages; the dollar changes and the 4.4x ratio are our arithmetic. Intraday, not closing, prices.
What Was Posted on X
The elections mean fewer or slower rollouts of new data centers, which makes the ones that already have a large DC install base more valuable: $AMZN, $GOOGL, $MSFT, $META. It also might mean CapEx becomes more gradual over the years for these companies (bad for semis)
The non-hyperscaler builders, like the neoclouds, will, on top of capital problems, now have further issues in giving more benefits to communities where data centers are built, which makes their ROI math even harder.
Rihard Jarc argues that fewer or slower data center rollouts make existing installed bases at Amazon, Alphabet, Microsoft and Meta more valuable, and that capex may become more gradual, which he calls bad for semis.
AI on the Ballot
Data Centers were on the ballot, and they lost in a major way; what does it mean if American opposition to building is permanent
Wall St Engine posted at 20:32Z on 4 November that President Trump told CNBC he can work with Democrats on ‘financial ideas’, which he said would be ‘mutually good’. The post relays the interview; we have not seen the full CNBC interview.
Update, 4 Nov, 21:02Z: Apple and Microsoft at the close.CNBC reported that “Hyperscalers and Apple surge after election results remove investor uncertainty”, writing that during trading Apple “climbed nearly 7%” and Microsoft “rose 4%”. It quoted David Kelly, chief global strategist at JPMorgan Asset Management, saying the divided government would lead to “moderation in growth, moderation in inflation, less uncertainty”.
Apple closed at $349.51 on 4 November, up $17.61 or 5.31% from $331.90, according to Nasdaq.com. With 14,532,182,000 shares outstanding on 23 October, per the cover of its 10-K filed on 3 November, that move added about $256 billion of market value, for about $5.08 trillion at the close, by our arithmetic.
A topic trending on X Explore read “Microsoft Hits $4.1 Trillion Market Cap on 7% Surge”. Microsoft closed at $555.80, up 6.68%, according to Nasdaq.com; at 7,426,194,552 shares outstanding on 22 October, per the cover of its 10-Q filed on 28 October, that gives a market value of about $4.13 trillion and about $259 billion added on the day, by our arithmetic. The intraday hyperscaler figures earlier in this piece are not added to these.
The Installed-Base Thesis
Rihard Jarc (@rihardjarc) posted on X at 15:06 UTC: “The elections mean fewer or slower rollouts of new data centers, which makes the ones that already have a large DC install base more valuable: $AMZN, $GOOGL, $MSFT, $META”. He added: “It also might mean CapEx becomes more gradual over the years for these companies (bad for semis)”.
Ben Thompson made a similar case in Stratechery on 4 November, under the heading “Hyperscalers Are Winners”. He wrote: “The downside for the neoclouds is upside for the hyperscalers, who are more likely to grow their share of AI compute than they were before”. He added that “investors may be pleased if the hyperscalers’ ambitions are somewhat curtailed”.
On chips, Thompson wrote under “Nvidia’s Mixed Bag” that “Nvidia will almost certainly be a short-term loser”, and also that “a constrained world makes Nvidia’s offerings more compelling”. His wider thesis: “The final constraint is political, and that may be the most imposing constraint of all”.
What the Ballot Showed
Thompson based his thesis partly on local votes, writing that “nearly every relevant local ballot measure opposing data centers won or is winning”. Our count of the data center ballot measures has limits leading in 40 of 41 votes on unofficial counts.
In The Boston Globe, Aaron Pressman wrote that the election “might have been even worse for big tech and their plans to spend trillions of dollars building AI data centers”.
He also wrote that “in dozens of local referendums around the country” voters “approved bans or restrictions on data centers”. He quoted analyst Dan Ives of Yorkville Ives calling the results a “big hurdle” for the proposed build-out, particularly in Texas.
As we read the moves, investors were pricing existing data center capacity differently from the chips that would fill new capacity: the companies that own large installed bases rose, while most of the companies that sell into new builds fell.
As we read Thompson and Jarc together, the argument is about where scarcity sits: if permission to build becomes the constraint, sites that already exist carry the value. That is their thesis; one trading day does not test it.
Ben Thompson, Stratechery
“The final constraint is political, and that may be the most imposing constraint of all.”
Three Implications
FOR HYPERSCALERS Thompson wrote that investors may be pleased if their ambitions are somewhat curtailed; by our arithmetic their shares added about $653 billion by 1:09 p.m. ET.
FOR CHIPMAKERS Jarc wrote that more gradual capex would be bad for semis; Micron, Nvidia, Broadcom and Intel were down at 1:09 p.m. ET, according to Nasdaq.com.
FOR DATA CENTER BUILDERS Earlier in the session, ten builders and power names were down an average of 11.8%, by our arithmetic.
The framework says: “This entire ecosystem has been shaped by constraints, bottlenecks, chokepoints, and scarcity at every layer”.
As we read it, 4 November tested a constraint the framework places at the infrastructure layers: if new data center sites become harder to permit, the value of existing sites and of the chips that fill them can move in opposite directions, as they did on the day. The framework predates the election and does not discuss it.
Update, 4 Nov, 18:19Z: A conservative AI policy group says AI policy shaped the result, Semafor reported. A post-midterms memo from Alliance for a Better Future, a conservative AI policy group, argues that Republicans need to recalibrate on AI before the next presidential election, Semafor reported. Semafor quotes the memo: “it’s impossible to ignore the strong impact Super Intelligence policy had on this election and how connected it is to economic concerns in particular”.
The memo also states, according to Semafor: “It would be political malpractice for Republicans not to course correct their SI policy positions and risk permanently ceding this issue to Democrats”. Semafor says the group recommends bills on chatbots and kids, codifying Trump’s voluntary AI accord into law and economic policies that help workers. The memo was shared with Semafor; we have not read it in full.
The figures are intraday prices, not closing prices. Share counts are from each company’s latest 10-Q or 10-K cover page, and Alphabet’s Class A move is applied to all three share classes, an approximation. Price moves on one day do not show why investors bought or sold, and none of the four hyperscalers has said whether the results change its capital spending plans.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
By our arithmetic, at 1:09 p.m. ET on 4 November Amazon, Microsoft, Alphabet and Meta had added about $653 billion of market value since the 3 November close, while Nvidia, Micron, Broadcom, Intel and AMD had lost about $149 billion net. Jarc and Thompson argued that slower data center building makes existing capacity more valuable; the market moves so far are consistent with that reading, but do not prove it.
A note on sourcing. Prices are Nasdaq.com real-time last-sale quotes at 1:09 p.m. ET (18:09 UTC) on 4 November 2026, against the 3 November close; they are intraday, not closing, prices. Share counts are from each company’s latest 10-Q or 10-K cover page on SEC EDGAR. Market value changes, totals and the 4.4x ratio are our arithmetic. Theses are attributed to Rihard Jarc (X), Ben Thompson (Stratechery) and Aaron Pressman (The Boston Globe).
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.
