Investor: AI reporters still don't grasp the basics of OpenAI and Anthropic's business models
firstadopter · x · 2026-09-13
Investor firstadopter says it's "crazy" that the average AI reporter still doesn't understand the basics of how OpenAI's and Anthropic's business models work, citing his year-old analysis:
- OpenAI's model splits into two parts: training new models requires enormous fixed investment, with each generation driving roughly 10x more spending, while inference on built models is extremely profitable.
- He restates his "AI is underhyped, not a bubble yet" case: reasonable Big Tech valuations with low leverage; we're in the early innings of the largest computing shift in decades (1994, not 1999); and every credible source reports overwhelming demand for AI compute with no overcapacity glut.
- He positions his coverage as facts and evidence over vibes and conflated narratives.
More from Venture
- Indie dev: ChatGPT and Claude now crawl his SaaS ~100 times a day — marclou · 2026-09-13
- Independent AI auditors wanted: backers offering advice, connections and funding — EricBuess · 2026-09-13
- Occam's razor: AI firms paused development to stop burning cash—expect ads and price hikes — OcelotCharacter1807 · 2026-09-13
- Polymarket puts 11% odds on an AI bubble burst by end of 2026 — Polymarket · 2026-09-13
- AI stocks may drop 10% Monday on Dario's blog post — or rally as market wakes up — devindkim · 2026-09-13
- Coefficient Giving launches Project Tailwind, a call for ambitious AI safety initiatives — lxrjl · 2026-09-13