Gary Marcus invokes steel-mill decline to argue neoclouds hinge on LLM profitability
GaryMarcus · x · 2026-09-21
Gary Marcus argues two points in a debate with investor Gavin Baker: the more profitable the LLM-providing business is as a whole, the better neoclouds will do — and if LLM returns on capital stay weak, the neocloud business will suffer.
He likens high-cost compute providers to America's collapsed steel industry, citing Baltimore's Sparrows Point mill (once the world's largest, 30,000+ workers at peak, closed for good in 2012).
Baker counters that the auto-driven rise of steel and oil consumption is the better analogy, since not all token-production inputs are commodities — and CoreWeave, Oracle and Nvidia valuations should logically be negatively correlated with model-layer margins.
Related event: Baker vs Marcus: Do Fat Model Margins Help or Hurt Neoclouds?(5 posts)→
More from Venture
- Chart: AI hits $55.6B annualized revenue in 2026, still trailing GLP-1 drugs — HealthcareAIGuy · 2026-11-12
- Nvidia scandal pushes TL20 portfolio to drop NVDA after 1,084% run since 2022 — TiernanRayTech · 2026-11-12
- 100 paid AI builds audited: 3 systems fetching $5k-$18k checks, chatbots only $1,100 — aryanXmahajan · 2026-11-12
- CoreWeave turns to customer prepayments and shorter contracts to fund AI growth — TechCrunch AI · 2026-11-12
- Bittensor Subnet NOVA Bets on Real Drug Assets to Prove Its AI Discovery Platform — markjeffrey · 2026-11-12
- Foreign Investors Pull $50B From Asian Equities in October on AI Cash-Burn and Regulation Fears — nordicinst · 2026-11-12