By Titik Puspa 
In a seismic development for the artificial intelligence landscape, AMI Labs, the ambitious new venture cofounded by AI luminary and Turing Prize winner Yann LeCun, has successfully closed a monumental funding round, raising an astonishing $1.03 billion at a pre-money valuation of $3.5 billion. This substantial investment underscores the immense potential and burgeoning interest in the nascent field of "world models" – AI systems designed to learn and understand the complexities of reality, moving beyond the purely linguistic data that has defined the current generative AI boom.
Yann LeCun, a pivotal figure in the advancement of deep learning, departed from his long-standing role at Meta to spearhead AMI Labs, a move that immediately positioned the startup at the forefront of AI research. The company’s core mission revolves around developing AI that possesses a grounded understanding of the physical world, a stark departure from the often abstract and disembodied nature of current large language models (LLMs). This focus on "learning from reality" is seen by many as the critical next step in AI’s evolution, promising more robust, reliable, and potentially safer AI applications.
Alexandre LeBrun, CEO of AMI Labs and a seasoned entrepreneur with a deep understanding of both AI and its real-world applications, articulated the company’s vision with a confident, albeit slightly playful, prediction. "My prediction is that ‘world models’ will be the next buzzword," LeBrun told TechCrunch, a smile evident in his tone. "In six months, every company will call itself a world model to raise funding." This statement, delivered with a hint of irony, highlights LeBrun’s conviction that while the term may become commoditized, AMI Labs’ fundamental approach is distinct and foundational.
The significance of AMI Labs’ focus on world models cannot be overstated, particularly in critical sectors like healthcare. LeBrun revealed that AMI Labs’ inaugural partnership will be with Nabla, a digital health startup of which he also serves as chairman. This strategic alliance is rooted in a shared recognition of the limitations inherent in current LLMs. As CEO of Nabla, LeBrun witnessed firsthand how the "hallucinations" or inaccuracies produced by LLMs could have severe, even life-threatening, consequences in a medical context. This realization, coupled with LeCun’s pioneering work on the Joint Embedding Predictive Architecture (JEPA) – a theoretical framework for building world models – forms the bedrock of AMI Labs’ research and development efforts.
LeBrun acknowledged that realizing the full potential of world models, particularly those based on JEPA, is a long-term endeavor. "AMI Labs is a very ambitious project, because it starts with fundamental research," he explained. "It’s not your typical applied AI startup that can release a product in three months, have revenue in six months and make $10 million in [annual recurring revenue] in 12 months." The journey from theoretical breakthrough to commercially viable product for world models is expected to span several years, demanding patience and sustained investment from both the company and its backers.
Despite the extended development timeline, the allure of world models has already attracted significant capital from investors. SpAI tial, another player in this emerging field, successfully raised a $13 million seed round, a notable sum for a European startup. More strikingly, Fei-Fei Li’s World Labs recently secured a staggering $1 billion, underscoring the burgeoning investor confidence in this AI paradigm. AMI Labs’ $1.03 billion raise places it firmly at the vanguard of this movement, exceeding earlier rumors that suggested a funding target of around €500 million.
The substantial funding achieved by AMI Labs can be attributed, in no small part, to the formidable team assembled. Beyond Yann LeCun’s esteemed chairmanship, the company boasts CEO Alexandre LeBrun’s proven entrepreneurial track record. Adding further weight to the leadership are Meta’s VP for Europe, Laurent Solly, as COO, and a cadre of high-profile researchers including Saining Xie as Chief Science Officer, Pascale Fung as Chief Research & Innovation Officer, and Michael Rabbat as VP of World Models. This concentration of talent and experience signals a deep commitment to pushing the boundaries of AI research.
The overwhelming investor interest provided AMI Labs with the strategic advantage of selecting partners whose expectations align with the company’s long-term vision. The funding round was co-led by prominent venture capital firms including Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions. The participation of numerous other funds, industry-tied backers, and notable individuals such as Tim and Rosemary Berners-Lee, Jim Breyer, Mark Cuban, Mark Leslie, Xavier Niel, and Eric Schmidt, further validates the profound belief in AMI Labs’ mission.
This significant capital infusion will provide AMI Labs with the crucial runway needed to address its two primary cost drivers: computational resources and top-tier talent acquisition. LeBrun emphasized a strategic approach to team building, prioritizing quality and expertise across four key global locations. The company will be headquartered in Paris, with significant operations in New York, where LeCun holds a professorship at NYU, and Montreal, the base for Michael Rabbat. A presence in Singapore is also planned, serving a dual purpose of tapping into Asia’s burgeoning AI talent pool and fostering proximity to potential future clients in the region.
While AMI Labs has no immediate plans for revenue generation, it is committed to early engagement with prospective users and partners. "We are developing world models that seek to understand the world, and you can’t do that locked up in a lab," LeBrun stated. "At some point, we need to put the model in a real-world situation with real data and real evaluations." This iterative approach, incorporating real-world feedback into the development process, is crucial for building truly effective world models.
The company intends to leverage partnerships for the deployment of its nascent models. Nabla stands as the first publicly disclosed partner, with many more expected to gain access to these early-stage technologies. LeBrun suggested that the strong interest from industrial players and potential collaborators during the investment round likely stems from this forward-looking strategy. The extensive list of backers further reinforces this collaborative vision, including not only lead investors and angels but also strategic corporate partners like NVIDIA, Samsung, Sea, Temasek, and Toyota Ventures. Prominent French entities such as Association Familiale Mulliez, Groupe Industriel Marcel Dassault, and Publicis Groupe also participated, alongside a host of other venture firms including Aglaé Lab, Alpha Intelligence Capital, Artémis, Bpifrance Digital Venture, New Legacy Ventures, SBVA, and ZEBOX Ventures.
While the path to commercial application for world models may be protracted, AMI Labs remains steadfast in its commitment to open research, a principle deeply ingrained in LeCun’s ethos. The company plans to regularly publish research papers detailing its progress. Furthermore, LeBrun affirmed that a significant portion of the developed code will be made open-source. This commitment to open research, a practice he noted is "increasingly rare," is driven by a fundamental belief that transparency and community collaboration accelerate progress. "We think things move faster when they’re open, and it’s in our best interest to build a community and a research ecosystem around us," LeBrun concluded, echoing a sentiment that has characterized some of the most impactful advancements in the field of artificial intelligence. The ambitious funding and clear strategic vision position AMI Labs as a pivotal player in shaping the future of AI, ushering in an era where machines not only understand language but also the world itself.