AMI Labs, a groundbreaking new venture cofounded by Turing Prize laureate Yann LeCun following his departure from Meta, has successfully closed an impressive funding round, securing $1.03 billion at a pre-money valuation of $3.5 billion. This substantial investment positions AMI Labs at the forefront of a burgeoning field: the development of "world models," a category of artificial intelligence designed to learn directly from reality rather than solely from linguistic data.

The company's core mission revolves around pioneering AI that can comprehend and interact with the physical world, setting it apart from the more prevalent generative AI landscape. Alexandre LeBrun, CEO of AMI Labs and a former CEO of digital health startup Nabla, anticipates that "world models" will soon become the next major buzzword in the AI industry. However, he emphasizes AMI Labs' distinct approach, aiming for a fundamental understanding of the real world, a goal he believes differentiates them from future contenders who might merely adopt the label for fundraising purposes.

The potential applications of world models are vast, with healthcare identified as an immediate area of focus. AMI Labs' initial strategic partnership is with Nabla, the digital health startup where LeBrun now serves as chairman. This collaboration stems from a shared realization between LeBrun and LeCun regarding the inherent limitations of large language models (LLMs), particularly their propensity for "hallucinations" – inaccuracies that could have dire, even life-threatening, consequences in sensitive fields like healthcare. AMI Labs aims to offer a robust alternative built upon the Joint Embedding Predictive Architecture (JEPA), a conceptual framework proposed by LeCun in 2022.

Despite the revolutionary potential, AMI Labs acknowledges the long-term nature of its endeavor. LeBrun characterizes the project as "very ambitious," rooted in fundamental research rather than immediate product deployment. Unlike typical applied AI startups that might achieve rapid revenue generation, world models require a significant investment of time, potentially years, to transition from theoretical concepts to viable commercial applications. This extensive timeline, however, has not deterred major investors, as evidenced by other significant funding rounds in the sector, such as SpAItial's $13 million seed round and Fei-Fei Li’s World Labs' recent $1 billion raise.

AMI Labs' funding round, which surpassed initial rumors of a €500 million target to reach approximately €890 million (around $1.03 billion), is a testament to the strength of its leadership and research team. The leadership roster includes not only LeCun as chairman and LeBrun as CEO but also Meta’s VP for Europe Laurent Solly as COO, high-profile researchers Saining Xie as Chief Science Officer, Pascale Fung as Chief Research & Innovation Officer, and Michael Rabbat as VP of World Models.

The robust interest from investors allowed AMI Labs to carefully select partners aligned with its long-term vision. The funding round was co-led by prominent firms such as Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions. Participation also came from a diverse group of other funds, industry-tied backers, and notable individuals including internet pioneers Tim and Rosemary Berners-Lee, venture capitalist Jim Breyer, entrepreneur Mark Cuban, tech veteran Mark Leslie, and prominent figures like Xavier Niel and Eric Schmidt. Additional strategic investors include NVIDIA, Samsung, Sea, Temasek, and Toyota Ventures, alongside significant French industrial players like Association Familiale Mulliez, Groupe Industriel Marcel Dassault, and Publicis Groupe, with further participation from Aglaé Lab, Alpha Intelligence Capital, Artémis, Bpifrance Digital Venture, New Legacy Ventures, SBVA, and ZEBOX Ventures.

This substantial capital infusion will primarily support AMI Labs' two main cost centers: advanced computing infrastructure and top-tier talent acquisition. LeBrun emphasizes a commitment to quality over quantity in building the team across four strategic global locations: Paris, serving as the company's headquarters; New York, where Yann LeCun teaches at NYU; Montreal, where Michael Rabbat is based; and Singapore, chosen for its strong AI talent pool and proximity to future Asian clients.

While AMI Labs currently has no immediate plans for revenue generation, it intends to engage proactively with prospective customers early in its development cycle. LeBrun explains this necessity, stating, "We are developing world models that seek to understand the world, and you can’t do that locked up in a lab. At some point, we need to put the model in a real-world situation with real data and real evaluations." This approach underscores the importance of real-world deployment and validation. Nabla, as the first disclosed partner, is anticipated to be among the early recipients of these models, with more industrial partnerships expected to emerge, explaining the strong interest from certain industrial players during the investment round.

A foundational principle for AMI Labs, consistent with LeCun’s long-standing beliefs, is a commitment to open science. The company plans to regularly publish research papers and make a significant portion of its code open source. LeBrun, who also previously worked at Meta’s AI research laboratory (FAIR), highlights that while open research is becoming increasingly rare, the founders believe it accelerates 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," he affirms, signaling AMI Labs' dedication to collaborative innovation and fostering a broader scientific community around world models.