Extropic founder claims 1000X more efficient AI inference could be worth trillions
beffjezos · x · 2026-09-05
Extropic founder Guillaume Verdon (beffjezos) predicts that 1000X greater inference energy efficiency could have a multi-trillion-dollar economic impact. He notes the company's first result a year ago was just MNIST — "the '80s of machine learning" — while today its models on a new hardware substrate already reach GPT-2 level, with algorithmic progress "blazing fast." The core bet: arbitrarily scalable models built on new hardware that avoids the wasteful matmuls GPUs perform.
More from Infra
- Is AI-Native Web Infrastructure a Real Category? Mapping TollBit, Exa, Parallel and Friends — Genzinvestor16180339 · 2026-09-05
- Extropic teases Z1T-0.1 as it races from GPT-2 toward GPT-3-level models — beffjezos · 2026-09-05
- Podcast Jokes About Scouting a 20GW Datacenter Site in Tahoe — chrisalbon · 2026-09-05
- Datacenters vs Shale Oil: Investors Wrong to Dismiss Selling Compute as a Commodity — abhiadesai · 2026-09-05
- Omnigent's Four-Line Defense Against Surprise LLM Bills: Routing, Budgets, Visibility — matei_zaharia · 2026-09-05
- GPT-6 Astra reportedly pretrained on 100,000+ GPUs at Stargate in Abilene, Texas — NinaDSchick · 2026-09-05