Advanced Machine Intelligence (AMI) Labs, a Paris-based AI startup co-founded by Turing Award winner Yann LeCun, closed a $1.03 billion seed funding round in early 2026, marking the largest seed round in European history, according to reporting by the Financial Times. The company, established earlier in 2026, reached a $3.5 billion pre-money valuation at close.
- AMI Labs raised $1.03 billion in seed funding — the largest seed round ever recorded in Europe
- The Paris-based company achieved a $3.5 billion pre-money valuation at close
- Investors include Nvidia, Temasek, Samsung, Bezos Expeditions, and former Google CEO Eric Schmidt
- AMI is building “world models” — AI systems designed to understand the physical world, maintain persistent memory, and support reasoning and planning, as an alternative to large language models
What Happened
AMI Labs completed its seed round with backing from a wide coalition of financial and strategic investors. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions. Additional participants include Nvidia, Temasek, Samsung, Toyota Ventures, Mark Cuban, and Eric Schmidt.
Yann LeCun — who shared the 2018 Turing Award for foundational contributions to deep learning and previously served as chief AI scientist at Meta — serves as Executive Chairman of AMI Labs. Alexandre LeBrun is CEO and Saining Xie serves as Chief Science Officer. Direct quotes from AMI’s leadership were not available in the source material at time of publication.
Why It Matters
The raise comes as LeCun has spent several years arguing publicly that large language models are fundamentally limited and do not truly understand the physical world. His position places him at odds with the prevailing direction of the major commercial AI labs, whose transformer-based language systems have dominated the market since 2022.
LeCun’s argument — that LLMs lack persistent memory, cannot reliably plan, and do not model physical causality — forms the intellectual basis for AMI’s research agenda. The $1.03 billion seed round, co-signed by investors including Nvidia and Temasek, suggests that critique has found serious financial backing beyond academic debate.
For Europe’s AI ecosystem, the scale of the raise is significant. European AI labs have historically struggled to compete with US and Chinese counterparts on the capital and compute needed for frontier research. A Paris-based company closing a billion-dollar seed round is without precedent on the continent and may attract further international investment and talent to the region.
Technical Details
AMI Labs is developing AI systems it describes as “world models” — architectures intended to build internal representations of the physical world, sustain persistent memory across interactions, and support structured reasoning and planning. This differs from current large language models, which generate outputs by statistically predicting sequences of tokens and do not, in LeCun’s framing, encode causal or physical structure.
According to AMI’s stated goals, these systems would be capable of understanding the real world, retaining memory over time, reasoning through problems, executing multi-step plans, and operating safely in open-ended environments. The company has identified robotics, manufacturing, wearable devices, and healthcare as its primary application domains — sectors where the shortcomings of LLM-based AI are most consequential in production settings.
As of this publication, AMI Labs has not released benchmark results, a technical paper, or a demonstration system that would allow external evaluation of its world model architecture.
Who’s Affected
Nabla, a clinical AI company, has been named AMI’s first strategic partner, with early access to the world model research currently underway. The partnership positions Nabla to integrate AMI’s work into clinical AI workflows before any public release, and gives AMI a real-world testing environment in a high-stakes domain where persistent memory and structured reasoning are operationally critical.
The breadth of AMI’s investor group — spanning European venture capital, US tech entrepreneurs, Asian corporate strategics including Samsung and Toyota Ventures, and sovereign wealth via Temasek — reflects cross-sector interest in world model research as both a scientific direction and a long-term commercial bet. Companies building products on top of large language model APIs may face future competition from world model-based systems if AMI’s research yields deployable results at scale.
What’s Next
AMI Labs has not disclosed a product roadmap, a research publication timeline, or specific technical milestones tied to this funding. The goals the company has outlined — persistent memory, physical world understanding, reliable planning, safe operation — remain unsolved research problems across the field. The $1.03 billion raise provides an extended runway to pursue this agenda without near-term commercial pressure.
The practical benchmark for AMI’s thesis will be whether its systems can demonstrate measurable advantages over existing LLM baselines on real-world tasks in robotics, healthcare, or manufacturing. Until published research is available, the $3.5 billion valuation reflects investor confidence in LeCun’s long-term research direction and leadership team rather than demonstrated product capability.
Related Reading
- Thailand’s Amity Raises $100 Million in Southeast Asia’s Largest Generative AI Round, Targets 2027 IPO
- Sauce Labs Launches AI-Driven Test Authoring Tool for Intent-Based Testing
- The Man Who Built Meta’s AI Just Left to Chase a $5 Billion Bet That LLMs Are a Dead End
- AI Megadeals Make Q1 2026 the Largest Venture Quarter on Record
