ThinkPatternGet the app
Perspective
BUSINESS · NOV 18, 2026

The adopter trade was finished before Wall Street named it

Wall Street's advice to rotate from AI infrastructure into "adopters" arrived this week as the last signature on a move the industry itself had been executing since February.

OpenAI cut its 2030 infrastructure plan from $1.4 trillion to about $600 billion in February, to line the buildout up with a $280 billion revenue forecast. [1] Nine months later, this week, Goldman Sachs said the industry's capital expenditure — money spent on chips, land, and data centers rather than day-to-day operations — would grow more slowly "for the first time." [2] The gap between those two dates is the whole trade. By September the labs had turned themselves into installation businesses. Microsoft created a Frontier company staffed with 6,000 forward-deployed engineers and a $2.5 billion commitment, to embed models in customers' systems; AWS put up $1 billion for the same push; OpenAI launched a deployment company of its own. [3] By mid-October the top models from Anthropic, OpenAI, Meta, and Google had reached near-parity — raw capability was no longer the differentiator. [4] October brought the funds' paperwork — the quarterly filings in which big investors disclose what they held three months earlier. Viking Global had sold its entire Apple stake; Appaloosa had exited Microsoft. The money moved into retail, home improvement, and pharma, with Eli Lilly the standout destination. [5] The lab resized in February, the installers deployed in September, the funds had filed their rotation by October. Every party executed before anyone advised. This week the trails crossed. Anthropic's S-1 — the filing that takes a company public — prices the record $1.4-to-$1.5 trillion IPO, backed by an $81 billion run rate, annualized revenue, and it makes its case on adoption, not intelligence. [6]

In the life sciences sector, where annual global pharmaceutical spend is expected to exceed $2 trillion in 2026, Claude is used in drug discovery and development by our customers, our biotechnology partners, and our own researchers — Anthropic

Eli Lilly, the funds' summer destination, became the lab's published exhibit: more than 85% of its 50,000 employees already use Claude. [7] Goldman's advice arrived this week — three weeks after JPMorgan's private bank made the same call [8] — the last signature on a stack that was already signed. [2] It told clients to move out of AI infrastructure and into adopters with physical businesses: ExxonMobil, GM, Waste Management. The same day, Nvidia reported revenue up 103% and guided its next quarter above consensus. [9] Demand is not rolling over. The rotation already has. The catch is already in the data. MIT's tracking found 95% of enterprise AI pilots produced no measurable effect on profit and loss. [10] And the S-1 carries its own exit clause — large customers may choose to build their own AI internally, and two customers already supply roughly a quarter of revenue. [6]

Large companies that are currently our customers may choose to develop their own AI capabilities internally rather than purchase our models. — Anthropic

One word closes it. "Adopter" is the name the desks gave the party the industry's own paperwork treats as a billable customer. The trade buys the billing layer, not an AI dividend.


Sources
  1. 1. OpenAI Cuts Infrastructure Spend to $600 Billion Ahead of IPO
  2. 2. Goldman Sachs Advises Shift Toward AI Adopter Stocks
  3. 3. AI Giants Commit Billions to Enterprise Deployment Services
  4. 4. Top AI Models Reach Parity as Market Focus Shifts
  5. 5. Hedge Funds Exit Major AI and Tech Stocks
  6. 6. Anthropic Targets $1.5 Trillion Valuation in Record-Breaking IPO
  7. 7. Anthropic Files for IPO With Pivot to Life Sciences
  8. 8. JPMorgan Private Bank Urges Shift Toward AI Adopters
  9. 9. Nvidia Corporation Revenue Surges 103 Percent to $115.8 Billion
  10. 10. Enterprise Agentic AI Pilots Fail Due to Data Gaps

Keep reading in the app

The full perspective, free in the app.