AI TL;DR
The Turing Award winner left Meta to build AI that understands physical reality. Here's everything we know about AMI Labs. This article explores key trends in AI, offering actionable insights and prompts to enhance your workflow. Read on to master these new tools.
Yann LeCun's AMI Labs: The $3.5 Billion "World Models" Startup
Yann LeCun—one of the three "godfathers of AI" and a Turing Award winner—has left Meta to start something new.
AMI Labs (Advanced Machine Intelligence Labs) is seeking a $3.5 billion valuation before even launching a product. Here's why investors are lining up.
Who Is Yann LeCun?
If you use any AI today, you're using LeCun's work.
His Contributions
- Invented convolutional neural networks (CNNs) — the foundation of computer vision
- Pioneer of deep learning alongside Geoffrey Hinton and Yoshua Bengio
- 2018 Turing Award winner (the "Nobel Prize of computing")
- Chief AI Scientist at Meta from 2013-2025
LeCun spent over a decade leading Meta's AI research, overseeing projects like LLaMA and the AI that powers Instagram and Facebook recommendations.
Why He Left Meta
LeCun has increasingly criticized the industry's obsession with Large Language Models (LLMs).
His argument: LLMs can generate text, but they don't understand the world.
He wants to build something different.
What Is AMI Labs Building?
"World Models"
AMI Labs is developing a new type of AI called world models—systems that understand physical reality, not just language.
Think about it:
- LLMs learn from text and can predict the next word
- World models learn from video/spatial data and understand physics, cause-and-effect, and common sense
Why It Matters
Current AI can write your emails but struggles with:
- Understanding why a ball falls when dropped
- Predicting how objects interact
- Reasoning about real-world scenarios it hasn't seen
World models could enable:
- Robots that actually work in unpredictable environments
- Self-driving cars that understand physics, not just lane markings
- AI assistants that can reason about the real world
The Team
Leadership
| Role | Person | Background |
|---|---|---|
| Executive Chairman | Yann LeCun | Turing Award, Meta AI |
| CEO | Alex LeBrun | Co-founder of Nabla (medical AI) |
| Team Member | Laurent Solly | Former Meta VP Europe |
Alex LeBrun
LeBrun is the founder of Nabla, a successful AI medical scribe startup. He'll remain chairman of Nabla while leading AMI Labs as CEO.
This background hints at AMI Labs' initial focus: healthcare.
The Funding
Valuation Hunt
AMI Labs is seeking to raise at a $3.5 billion (€3 billion) valuation—remarkable for a pre-product company.
Pre-Launch Round
Reports indicate the company is targeting €500 million ($586 million) in its first funding round.
Interested Investors
| Confirmed Interest | Status |
|---|---|
| Cathay Innovation | In discussions |
| Greycroft | In discussions |
| Hiro Capital | In discussions |
| 20VC | In discussions |
| Bpifrance | In discussions |
| Daphni | In discussions |
| HV Capital | In discussions |
| NVIDIA | In talks (oversubscribed) |
The NVIDIA interest is particularly notable—if the chip giant invests, it validates AMI Labs' approach to AI.
The Business Plan
Headquarters
AMI Labs is based in Paris, with planned offices in:
- New York
- Montreal
- Singapore
Target Industries
- Healthcare — Initial focus via Nabla partnership
- Industrial automation — Manufacturing, process control
- Robotics — Machines that understand physics
- Wearable devices — Next-gen ambient intelligence
Go-to-Market Strategy
AMI Labs plans to license its world model technology to partners rather than build end-user products directly.
Think NVIDIA selling chips vs. building PCs.
Timeline
LeCun stated AMI Labs expects to roll out AI models within approximately one year from January 2026.
Meta Connection
Will Meta Be a Customer?
There's speculation that Meta could become AMI Labs' first client.
LeCun maintains good relationships with Meta, and his world models could enhance Meta's future products—particularly AR glasses and robotics initiatives.
Academic Work
LeCun is keeping his academic affiliation with NYU, where he'll continue research that may feed into AMI Labs' development.
Open Source Philosophy
LeCun has championed open-source AI at Meta (see: LLaMA). He's indicated AMI Labs will follow a similar philosophy where appropriate.
The Bigger Picture
LeCun vs. The LLM Hype
LeCun has been vocal that the current approach to AI—training ever-larger language models—won't lead to true intelligence.
His critiques:
- "LLMs don't understand physics"
- "They hallucinate because they don't have a world model"
- "Scaling won't solve fundamental limitations"
AMI Labs is his bet that a different approach will work.
Competition
Other companies exploring similar territory:
- Google DeepMind — Gemini has world-knowledge elements
- Figure — Robotics with learned physics
- Tesla Optimus — Humanoid robots
- 1X Technologies — Robots with neural networks
AMI Labs' advantage: LeCun's unique expertise and track record.
Our Take
This is the most significant AI startup launch of 2026 so far.
When a Turing Award winner who invented core AI technologies starts a company at a $3.5B valuation to build something different from what everyone else is doing, you pay attention.
The bet is that LLMs are a local maximum—impressive but limited. World models might be the real path to capable AI.
If LeCun is right, AMI Labs could be foundational to the next decade of AI.
If he's wrong, it's still a fascinating experiment.
Either way: watch this space.
What do you think about world models vs. LLMs? Can physical understanding be trained into AI?