Kobeissi Letter: AI stocks fell over 13% in October while the “S&P 500 ex-AI” gained 8% to a record
An AI valuation pullback feeds into funding, compute pricing and enterprise purchasing — a macro signal for timing AI investments.
AI NEWS
Chinese and English are the same day. This page lists every item, each with the original link, the author, and the engagement.
Latest day
5 items for 2026-11-02, ranked by combined engagement.
Top 3
Kobeissi Letter: AI stocks fell over 13% in October while the “S&P 500 ex-AI” gained 8% to a record
An AI valuation pullback feeds into funding, compute pricing and enterprise purchasing — a macro signal for timing AI investments.
Read alongside October's AI-stock drop: real investment is still accelerating — the market dispute is about payback timing, not spending.
Frontier labs' cost of capital eventually shows up in API pricing and supply stability — a window into AI infrastructure debt risk.
All
An AI valuation pullback feeds into funding, compute pricing and enterprise purchasing — a macro signal for timing AI investments.
The Kobeissi Letter says a “stealth” bear market is under way: the Nasdaq 100 erased about $4 trillion in 25 days while the “S&P 500 Excluding AI” hit a record high on Friday. By its count AI stocks fell more than 13% in October as the ex-AI S&P 500 gained 8%. This is a market-commentary account's compilation; index definitions follow its chart. It points to money rotating out of the AI trade into other sectors, echoing recent doubts about returns on AI capex.
Read alongside October's AI-stock drop: real investment is still accelerating — the market dispute is about payback timing, not spending.
The Kobeissi Letter, citing US data, says private fixed investment in information-processing equipment and software grew at a 7.3% annualized rate in Q3 2026 to a record $1.77 trillion, the biggest quarterly increase since Q4 1983. At the peak of the 2000 dot-com bubble the metric grew 6.2%. By its count the category is up 61%, or about $671 billion, since Q4 2022, and has reached a record 5.4% of GDP versus the previous high of 4.5% in Q4 2000. The figures are its compilation and can be checked against BEA data; they show AI investment now exceeds dot-com-era intensity at the macro level.
Frontier labs' cost of capital eventually shows up in API pricing and supply stability — a window into AI infrastructure debt risk.
The Financial Times reports that Blackstone told investors in Anthropic's $63 billion chip-financing deal (Project North Star) that the company may not receive an investment-grade credit rating soon after its IPO, reversing earlier guidance of investment grade by year-end. Per secondary coverage of the FT story, Blackstone now sees a rating in the BB range or above (top of speculative grade) as more likely; the deal's $18 billion junior debt was rated BB+ (positive outlook) by KBRA and BB by Morningstar DBRS. The big three agencies have not issued ratings, and Anthropic and Blackstone declined to comment. Lacking investment grade would raise borrowing costs and limit buyers such as pensions and insurers.
Rapidly falling token prices are good news for AI users and bad news for compute-resale models.
Tom's Hardware reports that Darkbloom, a decentralized network selling spare Mac compute for inference, warned members to expect lower earnings. Per The Register, Anthropic's price per million tokens fell 77% in under five months, from $2.65 in May to $0.62 in October. Eigen Labs researcher Kydo says output per data-center cluster is up about 20x this year, leaving room to cut prices 90% and still earn 50% margins. The article estimates median monthly earnings of the top 50 operators at about $159.86, which could fall to roughly $1.23 a day after the price drop; Darkbloom says it will steer members toward high-demand models and other workloads such as CI/CD.
Which agent payment rails win will shape merchant fees and settlement, worth watching for businesses reliant on online payments.
X user alfongj writes that Stripe launched Link Balance for agents last week and that most takes miss the forest for the trees. He argues Stripe's real goal is to take over Visa and Mastercard's role as the world's payment network. This is the author's opinion; the article's details were not fully verified here, and Stripe has not framed it that way. It reflects how agent payments are becoming a new battleground among payment giants.