Yann LeCun’s AMI Labs raises $1.03 billion to build world models

AMI Labs, the groundbreaking venture founded by Turing Prize laureate Yann LeCun after his tenure at Meta, has successfully raised $1.03 billion at a pre-money valuation of $3.5 billion. The organization is focused on developing world models—AI systems that learn from the real world rather than solely from linguistic data. As the CEO of AMI Labs, Alexandre LeBrun envisions a future where 'world models' will dominate the tech landscape, potentially becoming the next industry buzzword. He confidently predicts that within six months, many companies will label themselves as world model developers to secure funding. However, LeBrun assures that AMI Labs stands apart, aiming to genuinely comprehend reality. The implications of this technology stretch into various sectors, particularly healthcare. AMI Labs has announced its first partnership with Nabla, a digital health startup for which LeBrun also serves as chairman. Having recognized the critical flaws in large language models (LLMs)—especially their potential to produce harmful inaccuracies—LeBrun is keen on establishing a practical alternative based on the Joint Embedding Predictive Architecture (JEPA) proposed by LeCun in 2022. LeBrun describes AMI Labs as an ambitious initiative, starting from foundational research rather than the typical applied AI approach where startups can quickly launch products and generate revenue. He acknowledges that developing world models may require years of work before reaching commercial viability. Despite the lengthy timeline, interest in world models is on the rise, with startups in this space securing substantial funding. For instance, SpAItial recently completed a $13 million seed round, and Fei-Fei Li's World Labs attracted a staggering $1 billion just last month. In this context, AMI Labs has exceeded its initial funding expectations, raising around €890 million, up from its original goal of €500 million. The team behind AMI Labs significantly contributes to its appeal, featuring LeCun as chairman, LeBrun's entrepreneurial background, and key figures like Meta’s VP for Europe, Laurent Solly, as COO, alongside other renowned researchers. The investment round, co-led by notable firms such as Cathay Innovation and Greycroft, included participation from several high-profile individuals and entities, ensuring a diverse and strategic investor base. This substantial influx of capital will support AMI Labs' primary expenditures: computational resources and talent acquisition. LeBrun aims to carefully build a specialized team across four key global locations: Paris, New York, Montreal, and Singapore, prioritizing quality over quantity. Although AMI Labs does not plan to generate revenue in the near term, it intends to connect with potential customers early in its development process. LeBrun emphasizes the importance of real-world testing for their models, stating, “We are developing world models that seek to understand the world, and you can’t do that locked up in a lab.” As partnerships unfold, Nabla will be the first to access these models, but certainly not the last. With backing from influential companies such as NVIDIA, Samsung, and Toyota Ventures, along with various French firms, AMI Labs is poised for significant advancements. While commercial applications may take time, the founders remain committed to open research, promising to publish papers and release code to foster a collaborative research community.