COMPANY

AMI Labs

companytopic-note

Overview

AMI Labs is Yann LeCun’s world-model startup, part of the H1 2026 world-model raise cluster tracked on 2026-07-14-AI-Digest alongside World Labs, Decart, Odyssey and 1X. The company closed a $1.03B seed round at $3.5B pre-money in March 2026 — reported at the time as Europe’s largest seed on record — co-led by Cathay, Greycroft, Hiro Capital, HV Capital and Bezos Expeditions, with Nvidia, Toyota Ventures, Temasek and Samsung as strategics. Michael Rabbat is a co-founder and is the operational face of the company at industry events.

Timeline

  • 2026-09-21-AI-Digest — On the Sept 20 All In conference panel, AMI Labs was represented by co-founder Michael Rabbat (not LeCun personally) alongside Fei-Fei Li’s World Labs; both companies declined to disclose commercialization and product plans, with Rabbat’s we'll talk about it when we're ready doing the load-bearing framing. Standing funding stack unchanged from the March 2026 seed. No new capex, term-sheet or partnership disclosed on stage — the secrecy on commercialization is itself the story, not a research-disclosure retreat. Corrective read: a subset of vision-forward world-model labs are going commercial-quiet (World Labs alongside AMI Labs), while DeepMind (Genie 3), Wayve (GAIA-3) and 1X keep publishing.

Key Developments

  1. March 2026 seed round ($1.03B at $3.5B pre-money): Europe’s largest seed on record at the time, co-led by Cathay, Greycroft, Hiro Capital, HV Capital and Bezos Expeditions with Nvidia, Toyota Ventures, Temasek and Samsung as strategics. Positions AMI Labs as a top-3 world-model raise cluster member alongside World Labs and 1X (roughly ~$4B primary-market valuation, per the July 14 cluster coverage).
  2. Commercial-quiet posture at All In (September 2026): Rabbat’s stage framing (we'll talk about it when we're ready) is the corpus’s first named data point that AMI Labs is holding commercialization and product plans back — pair with World Labs’ Aug 16 shipping-product artefact (2026-08-16-AI-Digest) as the first cluster member to publish, and today’s AMI panel as the first member to publicly decline on product terms.