Yann LeCun’s new AI startup, AMI Labs, is making waves in the artificial intelligence community. The company, founded by the renowned AI scientist after his departure from Meta, has officially announced its focus on developing world models – a foundational technology aiming to create AI systems with a deeper understanding of the real world. This announcement confirms speculation and positions AMI Labs as a key player in the rapidly evolving landscape of advanced machine intelligence.
AMI Labs launched its website this week, detailing its mission to “build intelligent systems that understand the real world” through the creation of these world models. The startup’s name, an acronym for Advanced Machine Intelligence, already hinted at this direction, but the official confirmation solidifies its place among leading AI research ventures attracting significant investment and talent.
The Rise of World Models in AI Development
Building foundational models that bridge the gap between artificial intelligence and the physical world has become a central focus in the field. This pursuit is attracting both top-tier scientists and substantial venture capital funding, even in the absence of immediate product releases. The approach represents a shift from solely language-based AI towards systems that can interact with and understand the complexities of the real world.
A direct competitor, World Labs, founded by AI pioneer Fei-Fei Li, quickly achieved unicorn status after its stealth launch. Following the release of its first product, Marble – a platform for generating physically accurate 3D environments – World Labs is reportedly seeking new funding at a $5 billion valuation. This success demonstrates the market’s appetite for this type of AI technology.
Funding and Leadership at AMI Labs
AMI Labs is also attracting significant investor interest. Reports indicate the startup may be raising funds at a $3.5 billion valuation, with potential investors including Cathay Innovation, Greycroft, and Hiro Capital, where LeCun currently serves as an advisor. Other firms reportedly considering investment include 20VC, Bpifrance, Daphni, and HV Capital.
However, a key detail regarding AMI Labs’ leadership structure has emerged: Yann LeCun holds the position of executive chairman, not CEO. That role is filled by Alex LeBrun, previously the co-founder and CEO of Nabla, a health AI startup based in Paris and New York. This division of responsibilities suggests a strategic focus on both research and practical application.
LeBrun’s appointment is a result of a partnership announced last December between Nabla and AMI Labs. In exchange for “privileged access” to AMI’s world models, Nabla’s board supported LeBrun’s transition to chief AI scientist and chairman, paving the way for his new role at AMI. This collaboration highlights the potential of world models in healthcare and other specialized fields.
A Contrarian Bet Against Large Language Models?
LeCun has been publicly critical of certain strategic decisions made by Meta under Mark Zuckerberg’s leadership. More broadly, analysts interpret AMI Labs as a deliberate move away from the current dominance of large language models (LLMs). LeCun has voiced concerns about the limitations of LLMs, particularly their tendency towards “hallucinations” – generating inaccurate or nonsensical information – a critical issue in sensitive areas like medicine.
AMI Labs aims to address these limitations by creating AI systems that possess persistent memory, reasoning capabilities, and controllability. Unlike generative AI approaches, the startup promises its technology will be well-suited for handling unpredictable real-world data, such as sensor input. The company plans to license its technology to industry partners for practical applications, while also contributing to the broader AI research community through open publications and open-source initiatives.
The startup’s mission statement emphasizes its focus on applications where reliability and safety are paramount, including industrial process control, automation, robotics, and healthcare. AMI Labs believes that true intelligence begins not with language, but with understanding the world around us. This focus on embodied intelligence represents a significant divergence from the current trend of LLM development.
AMI Labs will maintain a global presence with offices in Montreal, New York, and Singapore, but has chosen Paris as its headquarters. This decision was welcomed by French President Emmanuel Macron, who expressed his support for LeCun and the company’s commitment to France. The move further solidifies Paris’s growing reputation as a hub for artificial intelligence research and development, joining existing players like Mistral AI and Meta’s FAIR lab.
The next steps for AMI Labs will likely involve securing its funding round and accelerating the development of its world models. The company’s progress, and the broader adoption of this technology, will be closely watched by the AI community and investors alike. The success of AMI Labs could signal a significant shift in the direction of AI research, moving beyond language-centric models towards systems that are more deeply integrated with the physical world.
