Yann LeCun’s AMI secures €890 million for next‑gen AI
AI pioneer Yann LeCun, one of the key architects behind modern deep learning, has launched a new venture, AMI, and raised an impressive €890 million to develop so‑called world model AI systems. The funding round signals growing investor appetite for alternatives to today’s dominant large language models (LLMs).
Beyond large language models
Unlike mainstream AI products that rely heavily on text‑based LLMs, AMI aims to build AI that can understand and predict the physical and social world more like humans do. World models are AI architectures that integrate perception, memory, and reasoning to form internal representations of how the world works, enabling richer planning and decision‑making.
LeCun has long argued that current generative AI tools are powerful but fundamentally limited, as they primarily model language rather than reality itself. With AMI, he intends to push toward systems that can learn from video, sensor data, and interaction, not just text, and that can act autonomously in complex environments.
Strategic funding for ambitious research
The €890 million raise puts AMI among the best‑funded AI startups worldwide. While specific investors were not disclosed in the source text, the scale of the round suggests participation from major venture capital and institutional investors seeking exposure to foundational AI research.
The capital will be used to expand AMI’s research teams, build large‑scale computing infrastructure, and experiment with new AI architectures that combine self‑supervised learning, reinforcement learning, and robotics. The company is expected to compete with leading labs focused on general‑purpose AI systems, while pursuing a more science‑driven approach to intelligence.
Implications for the global AI race
The launch of AMI reinforces Europe’s ambition to remain relevant in the global AI race, traditionally dominated by US and Chinese tech giants. By backing LeCun, investors are betting that world model architectures could define the next era of AI, moving beyond chatbots toward agents that can understand, predict, and safely operate in the real world.
For the broader ecosystem, AMI’s work could influence how future AI safety, regulation, and industrial applications are shaped, from autonomous robots and virtual assistants to scientific discovery tools.