Yann LeCun launches AMI Labs: the next revolution in artificial intelligence

In summary

• Yann LeCun launches AMI Labs, a startup dedicated to the next generation of AI
• The company raised over one billion dollars in its first funding round
• The goal is to create systems capable of understanding the real world
• Models are trained on videos and physical data
• These systems could transform robotics, scientific research and industry

French researcher Yann LeCun, deep learning pioneer and former Chief AI Scientist at Meta, believes the next AI revolution will not come from current generative models, but from systems capable of understanding and anticipating the real world.

In a recent interview on France Inter, he explains that current models — although impressive — remain limited. To reach the next level, artificial intelligence must learn to reason about the consequences of actions in the physical environment.

This is precisely the goal of his new company, AMI Labs (Advanced Machine Intelligence).

A historic funding round of over one billion dollars

Hardly launched, AMI Labs is already positioning itself as one of the most ambitious AI startups in the world.

The company announced raising $1.03 billion (approximately €890 million) in its first funding round. This round values the startup at $3.5 billion.

This is one of the largest Series A funding rounds in the history of artificial intelligence. Only Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati, exceeded this amount with $2 billion.

The round brings together around twenty international investors, including:

• Nvidia
• Bezos Expeditions (Jeff Bezos's fund)
• The Muliez family
• The Dassault family
• Artémis (the Pinault family holding)
• Xavier Niel

Notably, Meta — LeCun's former employer — did not participate. However, he considers the company a potential partner for scientific collaborations.

Why current AI does not understand the real world

Unlike current language models that learn primarily from text and static images, Yann LeCun points to a fundamental limitation: these systems do not understand the physical world. They are effective at generating content but cannot anticipate the consequences of actions in real environments.

This gap between text generation and world understanding is at the heart of the AMI Labs project.

The technical approach: world models

Yann LeCun proposes an approach based on what he calls world models — models that learn to build an internal representation of the world, similar to that used by humans.

These systems integrate several fundamental capabilities:

• Perception: understanding what is happening in the environment
• Memory: retaining past experiences
• Reasoning: drawing logical conclusions
• Planning: imagining sequences of actions before executing them

AI learning from videos

To train these world models, systems are fed large quantities of videos. The idea is simple: the real world is dynamic, and a video contains enormous information about how objects interact in space and time.

By observing millions of hours of videos, an AI system can learn:

• How objects move
• How actions produce consequences
• How situations evolve over time

For example, a model trained on videos can learn that a falling cup will likely break, or that an object sliding off a table will fall to the floor. This ability to anticipate how a situation will evolve is central to human intelligence.

Systems capable of reasoning and planning

In this approach, AI is no longer just a system that generates text. It becomes a system that predicts future states of the world and simulates actions before executing them.

These capabilities open major perspectives:

• Simulating scientific experiments
• Virtually testing molecules
• Planning actions for a robot
• Anticipating accidents in a real environment

The medium-term goal: create universal intelligent systems reaching intelligence comparable to that of an adult human in certain cognitive tasks.

Future applications: robotics, science and industry

AMI Labs targets several sectors to be transformed by this new generation of AI:

• Robotics: robots capable of naturally interacting with their environment, anticipating object falls, planning action sequences
• Autonomous driving: real-time planning in complex environments
• Scientific research: experiment simulation, drug discovery
• Industry: production chain management, autonomous industrial systems

Although founded by a French researcher, AMI Labs positions itself as an international company with offices in Paris, New York, Montreal and Singapore.

A world pioneer in artificial intelligence

Yann LeCun is considered one of the most influential scientists in AI. His work on convolutional neural networks revolutionized image recognition. In 2018, he received the Turing Award, often considered the "Nobel Prize of Computer Science".

The ethical stakes of artificial intelligence

The development of ever more powerful AI systems raises important ethical questions. LeCun is clear on this point:

> "The decision about what is the best use of AI for society should not be in the hands of researchers or companies, but in the hands of society and its democratic institutions."

FAQ – AMI Labs and Yann LeCun's vision

What is the AMI Labs startup?

AMI Labs (Advanced Machine Intelligence) is a startup founded by French researcher Yann LeCun. Its goal is to develop a new generation of artificial intelligence capable of understanding the physical world, reasoning and planning complex actions.

Why does Yann LeCun criticize current AI models?

According to Yann LeCun, current models like LLMs learn primarily from text and have no real understanding of the real world. They are effective at generating content but cannot anticipate the consequences of actions in the physical environment.

How does AMI Labs train its AI models?

The approach relies on learning from videos. By analyzing millions of hours of video sequences, models learn how objects interact in the real world and can predict how a situation will evolve.

How large is AMI Labs' funding round?

AMI Labs raised over one billion dollars in its first funding round, valuing the company at approximately $3.5 billion.

In which fields could this technology be used?

Systems developed by AMI Labs could transform several sectors: robotics, autonomous driving, industry, scientific research and drug discovery.

🎧 Listen to the interview

Find the full interview with Yann LeCun on France Inter: L'invité de 7h50 – Yann LeCun – France Inter, March 10 2026

At Busony, we deploy conversational AI agents and MCP systems today to connect Claude to your WordPress or WooCommerce site. Discover our services: AI Agents · WordPress AI · WooCommerce + AI