Summary

AMI Labs (Artificial Machine Intelligence Labs), led by AI pioneer Yann LeCun, has completed a $1.03 billion seed round — the largest in European history — at a pre-money valuation of $3.5 billion. Backed by Bezos Expeditions, Nvidia, Samsung, and Temasek, the company is embarking on the development of a "world model" centered on JEPA (Joint Embedding Predictive Architecture). This raise sits at the heart of a historic week in which over $6 billion flowed into AI in just three days, symbolizing that physical AI and vibe coding are approaching "escape velocity."

Text

On March 19, 2026, AMI Labs (Artificial Machine Intelligence Labs), founded by Yann LeCun, announced a seed round of $1.03 billion. With a pre-money valuation of $3.5 billion, the round broke the historical record for a seed round among European startups. Companies that have raised this level of funding at the seed stage are exceedingly rare on a global scale.

Betting Against LLMs — The JEPA Architecture Wager

LeCun has long maintained a publicly critical stance toward the token-prediction-based approach adopted by today's large language models (LLMs). The transformer-based LLMs in which OpenAI, Anthropic, and Google compete rely on predicting the next token in a sequence of text, but LeCun argues that this approach can never achieve "true intelligence."

The JEPA (Joint Embedding Predictive Architecture) adopted by AMI Labs is an architecture that learns the "abstract representations" of the world itself, rather than predicting tokens one by one. This is said to more closely resemble the way humans understand the physical world. Humans do not understand the world by predicting text one word at a time; they intuitively grasp the structure and dynamics of their entire environment. JEPA is an attempt to implement this intuitive understanding in AI.

Leadership — Merging Research and Commercialization

The leadership team at AMI Labs combines top AI researchers with operators who have proven business track records. LeCun himself oversees the direction of research as Chairman. Alexandre LeBrun has been appointed CEO. LeBrun is the former CEO of digital health company Nabla and brings extensive experience in the commercialization of AI technology. COO Laurent Solly led Meta's European operations as VP of Europe, with a strong track record in large-scale organizational management and business development.

On the research side, Saining Xie has been appointed CSO (Chief Science Officer), Pascale Fung as CRIO (Chief Responsible Innovation Officer), and Michael Rabbat as VP of World Models. Xie is a globally recognized researcher in computer vision, Fung is an authority on responsible AI development, and Rabbat has led world model research at Meta AI.

Investor Lineup and Strategic Implications

Investors include venture funds such as Cathay Innovation, Greycroft, Hiro Capital, and HV Capital, alongside Bezos Expeditions (Jeff Bezos's personal investment fund), Nvidia, Samsung, and Temasek (Singapore's government investment fund). Nvidia's participation is particularly noteworthy: given that the JEPA architecture requires large-scale computational resources, it carries strategic significance from the perspective of securing GPU supply. Samsung's involvement reflects its ambition to integrate AI into smartphones and smart home appliances.

Target Markets and Global Expansion

The markets AMI Labs is targeting span six domains: manufacturing, aerospace, biomedical, pharmaceuticals, robotics, and healthcare — all areas that demand an understanding of the physical world, which is precisely where today's text-centric LLMs struggle. Nabla, the digital health company where CEO LeBrun previously worked, was announced as the first partner. The collaboration with Nabla signals AMI Labs's intent to simultaneously pursue its technical ambition of demonstrating world models in the medical field and the pace of commercialization.

Offices will operate across four locations: Paris (headquarters), New York, Montreal, and Singapore. The plan is to build a global research and business organization from day one, despite being a European-born company.

The Centerpiece of a Historic Week

The AMI Labs raise was not an isolated event. That same week saw Mind Robotics ($500 million), Rhoda AI ($450 million), Replit ($400 million), Legora ($550 million), and Sunday ($165 million) — over $6 billion poured into AI in just three days. It was a week in which the market declared that two currents — Physical AI (robotics) and vibe coding (AI-native development) — are approaching escape velocity. Unlike the LLM-concentrated investment of 2023–2024, the fact that capital is now distributed across all layers — foundation models, the application layer, physical AI, and infrastructure — speaks to the maturation of the AI industry.

Impact on the Industry

The emergence of AMI Labs has thrust a concrete alternative into the market in response to a fundamental question about the direction of AI research — "Can scaling LLMs alone achieve general intelligence?" LeCun is one of the most influential AI researchers in the world, and the fact that his position has been backed by funding on the scale of $1 billion (approximately ¥150 billion) has the potential to influence how the entire industry allocates research investment.

It is also symbolic for Europe's AI ecosystem. Following Mistral AI's revitalization of France's AI industry, AMI Labs has demonstrated at an even larger scale that Europe can stand at the forefront of AI research. An approach distinct from the transformer-based competition led by American players is poised to emerge from Europe.

The ripple effects on the robotics market are also significant. If world models that understand the physical world are put to practical use, the intelligence level of industrial robots will improve dramatically, potentially bringing transformation to a wide range of industries — manufacturing, logistics, agriculture, construction, and even healthcare. This historic week, in which $6 billion (approximately ¥900 billion) was deployed over three days, may in later years be looked back upon as "the moment the Cambrian explosion of physical AI began."

Reference information: AMI Labs founding announcement, Yann LeCun's JEPA architecture paper, PitchBook AI investment tracker, CB Insights funding data, European startup investment trends