Yann LeCun Raises $1B to Challenge AI Industry’s Path
▼ Summary
– Yann LeCun, a Turing Award winner, left Meta four months ago believing large language models (LLMs) are fundamentally limited and not true intelligence.
– He has now secured $1.03 billion in seed funding for his new company, AMI, which is Europe’s largest-ever seed round and values the startup at $3.5 billion.
– AMI aims to build “world models” using an alternative AI framework called JEPA, which focuses on learning abstract representations of how the world works rather than predicting words.
– The company is positioned as a European counterweight to American and Chinese AI giants, with its headquarters in Paris and a founding team largely drawn from Meta’s AI research division.
– Despite having no product or revenue, the massive funding demonstrates investor confidence in LeCun’s vision and research record, though the success of world models remains a long-term, open question.
A new European artificial intelligence venture has secured an unprecedented level of funding, signaling a major bet on an alternative path for the field. Advanced Machine Intelligence Labs (AMI), founded by Turing Award laureate Yann LeCun, announced a staggering $1.03 billion seed investment. This capital, Europe’s largest seed round on record, backs LeCun’s long-stated conviction that the industry’s prevailing focus on large language models (LLMs) is fundamentally limited. The company aims to pioneer a different class of AI known as world models, positioning itself as a European challenger to American and Chinese tech dominance in artificial intelligence.
LeCun departed his role as Meta’s chief AI scientist in late 2025, having grown disillusioned with the trajectory of mainstream AI development. He has consistently argued that LLMs, while impressive in generating fluent text, represent a form of statistical pattern matching rather than genuine intelligence. These models learn by predicting the next word in a sequence, a process LeCun believes cannot lead to true understanding or reasoning about the physical world. His new venture, AMI, is founded on the principle that a more robust foundation for machine intelligence is required.
The substantial funding round was co-led by five investment firms: Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Jeff Bezos’s personal investment vehicle, Bezos Expeditions. A broad consortium of strategic and financial backers also participated, including industry giants like Nvidia, Toyota, and Samsung, alongside prominent venture capital firms and individual investors such as Tim Berners-Lee and former Google CEO Eric Schmidt. The company’s valuation was set at $3.5 billion prior to the investment, reflecting immense confidence in LeCun’s vision despite the early, purely research-focused stage of the company.
AMI’s operational headquarters will be in Paris, with plans for offices in New York, Montreal, and Singapore. LeCun, who retains a professorship at New York University, will serve as executive chairman. Day-to-day leadership falls to CEO Alexandre LeBrun, a French entrepreneur with a background in medical AI. The founding team is heavily populated by LeCun’s former Meta colleagues, including former research directors and executives, creating a deep bench of experienced AI research talent.
The technical core of AMI’s mission is the development of world models. This approach moves beyond next-word prediction to systems that learn abstract, internal representations of how the world functions. LeCun’s proposed framework, called the Joint Embedding Predictive Architecture (JEPA), focuses on understanding the underlying relationships and dynamics of environments, ignoring unpredictable surface details. The goal is to create AI that comprehends reality through a form of embodied experience, similar to humans and animals, rather than through language alone.
LeCun has outlined an ambitious timeline for the company. He anticipates beginning discussions with corporate partners within one to two years, with the goal of deploying “fairly universal intelligent systems” across various domains within three to five years. The explicit European, and particularly French, identity of AMI is a strategic choice. LeCun has positioned the lab as a direct counterweight to the concentrated power of U.S. and Chinese AI giants, stating it is “one of the few frontier AI labs that are neither Chinese nor American.”
The road ahead, however, is long and uncertain. AMI currently has no commercial product, no revenue, and plans to focus entirely on research and development for its first year. World models represent a profound scientific challenge, far removed from the rapid product cycles of typical AI startups. The monumental seed investment is ultimately a wager on LeCun’s unparalleled research pedigree and his persistent critique of current AI architectures. While the industry acknowledges the limitations of LLMs, the critical question remains whether LeCun’s alternative vision can translate from compelling theory into practical, world-changing technology.
(Source: The Next Web)
