Ex-Meta AI chief Yann LeCun’s AMI raises $1.03 billion for alternative AI approach

Summary

Advanced Machine Intelligence (AMI), the startup founded by former Meta AI chief Yann LeCun, has raised $1.03 billion as investors back an alternative path to artificial intelligence focused on reasoning and real-world understanding rather than large language models.

SAN FRANCISCO, March 10, 2026 — Advanced Machine Intelligence (AMI), the startup founded by former Meta Platforms chief AI scientist Yann LeCun, said on Tuesday it had raised $1.03 billion in its first major funding round, valuing the company at $3.5 billion pre-money.

The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, and HV Capital, with participation from Bezos Expeditions, the investment firm of Jeff Bezos.

The investment highlights growing venture interest in approaches to artificial intelligence that go beyond large language models (LLMs).

Moving beyond “next-word” prediction

LeCun, a pioneer of deep learning and co-recipient of the 2018 Turing Award, has frequently argued that current LLM-based systems lack genuine reasoning ability and real-world understanding.

Speaking to Reuters, he said systems built primarily on predicting the next word or pixel cannot alone produce broadly capable intelligent agents.

AMI instead aims to develop AI built on so-called “world models,” which simulate how environments function and allow systems to reason, plan and anticipate outcomes.

Industrial and consumer ambitions

Unlike many startups focused on conversational AI tools, AMI is targeting industries that manage complex physical and biological systems.

Near-term applications include:

• Automotive and aerospace: Advanced autonomy and navigation planning.
• Manufacturing: Optimization of high-precision robotic workflows.
• Biomedical research: Modeling complex biological systems for drug discovery.

LeCun said the long-term goal is to enable domestic robots capable of navigating real-world environments safely and autonomously.

He also confirmed discussions with Meta over potential use of AMI’s reasoning systems in future versions of Ray-Ban Meta smart glasses.

A philosophical break from Meta

LeCun’s departure from Meta in late 2025 followed the company’s internal shift toward large-scale generative AI efforts under leadership including Alexandr Wang.

Having founded Facebook AI Research (FAIR) in 2013, LeCun’s pivot to AMI marks both a technical and philosophical departure from the industry’s current emphasis on scaling LLMs.

Why this matters

• A growing AGI divide: The funding signals investor belief that alternatives to LLM scaling may shape the next generation of AI.
• Real-world automation potential: World-model systems could unlock robotics applications in unpredictable environments such as logistics, construction and healthcare.
• Industry competition shift: AMI’s rise intensifies competition among infrastructure-heavy AI firms pursuing different technological paths.
• Strategic relevance for Meta: Ongoing collaboration discussions suggest continued overlap between LeCun’s vision and Meta’s hardware ecosystem.

FAQs

Q1. What are “world models”?

They are AI systems designed to simulate how environments work, enabling reasoning, planning and prediction beyond language tasks.

Q2. Why did Yann LeCun leave Meta?

His departure followed a strategic shift toward generative AI at Meta, differing from his emphasis on reasoning-based systems.

Q3. Will AMI compete with OpenAI?

Partly. AMI focuses more on intelligent agents and industrial applications than consumer-facing generative AI tools.