Why model-layer margins should be negatively correlated with CoreWeave, Oracle and Nvidia
GavinSBaker · x · 2026-09-21
Investor Gavin S. Baker pushes back on the "cars boosted steel and oil" analogy for AI compute: token production inputs aren't all commodities.
His core argument: CoreWeave, Oracle and Nvidia valuations should logically be negatively correlated with model-layer margins. Token production cost is effectively those companies' revenue; the higher the margin OpenAI/Anthropic add on top, the fewer tokens get produced and the less revenue flows to infra. He adds that an OpenAI/Anthropic monopsony would be bad for everyone in AI infrastructure.
Related event: Baker vs Marcus: Do Fat Model Margins Help or Hurt Neoclouds?(5 posts)→
More from Venture
- AI assistant startup Town tops The Information's 50 Most Promising Startups at $892M valuation — soleio · 2026-11-12
- OpenAI CFO Sarah Friar says OpenAI has 100 million subscribers — morqon · 2026-11-12
- Polymarket odds put Chinese AI lab at #1 by year-end at just 6% — Polymarket · 2026-11-12
- Rox raises $270M Series B at $1.3B valuation, 12 months after launch — Konstantine · 2026-11-12
- TSMC to invest $1B in OpenAI and Anthropic, fueling circular financing concerns — Ars Technica AI · 2026-11-12
- Memory stocks rebound as CoreWeave forecast signals demand: SanDisk +16.75% — fallingdowndizzyvr · 2026-11-12