AMI Labs' work is AI for the physical world and is not language-related.

True if, as of 2026-10-01, AMI Labs' stated research focus and published work are on AI for the physical world (world models, learning from video, sensors or physical interaction, for example JEPA-style architectures) and do not centre on language models. Mostly true if the focus is the physical world but some secondary component involves language, for example a language interface. Mixed if language-related work is a substantial part of what it does alongside physical-world AI. Mostly false if its main products or research are largely language-based, with physical-world AI only an aspiration. False if AMI Labs' work is mainly language modelling.

Set before the research began.

1. E1 AMI Labs · HansaIndex HansaIndex · 9 October 2026

   “AMI Labs is a company that develops world models for robotics. It focuses on creating models that enhance the capabilities of robotic systems, rather than language processing.”

2. E2 AI pioneer Yann LeCun discusses new frontiers in the field at Brown lecture The Brown Daily Herald · 2 April 2026

   “LeCun’s company, AMI Labs, is an AI startup that is developing the concept of “world models,” which are AI models able to understand the physical world because they are trained with sensory data, as opposed to language-based models, which are trained on human-produced texts.”

3. E3 Yann LeCun, AMI Labs | Evident AI Symposium 2026 — Yann LeCun CEOInterviews.AI · 27 October 2026

   “Yann LeCun, co-founder and executive chairman of AMI Labs, discussed the company's focus on building AI for the physical world, distinguishing it from LLMs.”

4. E4 Yann LeCun: The deep learning pioneer who walked away from the LLM race keep-thinking.com · 23 January 2026

   “The company says it will publish papers and open-source code as it goes, a deliberate continuation of the FAIR model, and it is candid that commercial applications of world models may take years.”

Two independent sources describe AMI Labs as working on world models for the physical world rather than on language. E2 (April 2026) says its models are trained on sensory data rather than human-produced text, and E1 (dated October 9, 2026, after the judgment date) says it builds world models for robotics rather than language processing. E3 (dated October 27, 2026, after the judgment date) also describes the focus as physical-world AI as distinct from LLMs. None of the evidence points to a language-centred component.