Europe is backing a new wave of AI research that aims to move beyond today’s dominant large language models and build systems that actually understand the physical world they operate in. At the center of this push is Advanced Machine Intelligence (AMI), a new lab co‑founded by renowned computer scientist Yann LeCun, formerly chief AI scientist at Meta and a Turing Award laureate. AMI has just secured about 1.03 billion dollars in seed funding, which is not only Europe’s largest seed round to date, but also a clear signal that investors believe in an alternative path to artificial intelligence. The company is headquartered in Paris, with additional hubs planned in New York, Montreal and Singapore, underlining the ambition to build a truly global research operation rooted in Europe’s tech ecosystem.
LeCun has long argued that current AI systems based mainly on text prediction, such as popular chatbots, are inherently limited for deeper reasoning and real‑world decision‑making. Instead, AMI wants to develop so‑called “world models” that can learn from video, interaction and sensory data, building internal representations of objects, space, causality and physical constraints. In practice, that means models designed not just to generate fluent text, but to guide robots, vehicles or industrial systems that must navigate complex environments safely and autonomously. The startup is already working on an AI system tentatively named AMI Video, aimed at understanding and predicting what happens in real‑world scenes, with applications in robotics, manufacturing, transportation and wearables.
The funding round is backed by a mix of US, European and Asian investors, including Bezos Expeditions, Nvidia, Temasek, Cathay Innovation, Daphni, HV Capital, Hiro Capital and SBVA. This diversity of backers suggests that AMI’s research is seen as strategically important well beyond Europe. At the same time, it fits into a broader European trend of building homegrown AI labs and infrastructure, alongside initiatives like the French open‑source‑oriented lab Kyutai in Paris, which has raised around 300 million euros for foundational research. Together, these efforts show Europe is not only trying to regulate AI, but also to shape its technical foundations by investing heavily in new architectures that could complement or even rival the current generation of ChatGPT‑style systems.
