Yann LeCun unveils his new startup Advanced Machine Intelligence (AMI Labs) -- and raises $1.03B
The AI pioneer's new startup bets $1B that modeling reality, not just text, is the next breakthrough.
Yann LeCun, a foundational figure in modern AI, has launched a new venture called Advanced Machine Intelligence (AMI Labs) alongside co-founder Alexandre LeBrun, former CEO of Nabla and founder of Wit.ai. The startup has secured a staggering $1.03 billion in funding from a who's-who of tech investors, including NVIDIA, Samsung, Bezos Expeditions, and Eric Schmidt. Their mission is to move beyond the limitations of current large language models (LLMs) by building AI systems that model the physical world, not just language patterns.
AMI Labs is betting on LeCun's Joint Embedding Predictive Architecture (JEPA) as the technical foundation for creating these 'world models.' The core thesis is that LLMs' tendency to hallucinate represents a fundamental ceiling, especially for critical applications like healthcare. Instead of predicting the next word, their AI would learn an internal model of how the world works. The company is explicitly focused on long-term, open-source fundamental research, with no immediate product plans, framing this as a 5-10 year endeavor.
This massive funding round for a pure research lab signals a major shift in AI investment priorities. Backers are placing a billion-dollar bet that the next paradigm in AI will require a deeper, more grounded understanding of reality than today's text-trained models can provide. LeBrun predicts 'world models' will become the industry's next major buzzword, and AMI Labs aims to be the entity that delivers the substantive research behind the hype.
- Raised $1.03B from elite investors including NVIDIA, Samsung, and Bezos Expeditions for pure AI research.
- Building 'world models' using LeCun's JEPA architecture to move beyond hallucination-prone LLMs by modeling physical reality.
- Open-source, long-term (5-10 year) research play with no short-term product, led by an all-star team of AI researchers.
Why It Matters
This massive bet could redirect the entire field of AI from language models toward systems that genuinely understand and reason about the world.