Paris, France – A new artificial intelligence startup, Advanced Machine Intelligence (AMI), has secured over $1 billion in funding, signaling a significant bet on a different path for AI development. Founded by Yann LeCun, the renowned AI researcher and former chief AI scientist at Meta, AMI aims to build AI systems that move beyond the current focus on large language models (LLMs) and instead develop “world models” – AI capable of understanding the physical world, reasoning, and planning complex actions. The funding round, announced Tuesday, values the company at $3.5 billion, reflecting investor confidence in LeCun’s vision.
This substantial investment underscores a growing debate within the AI community regarding the optimal route to achieving true artificial general intelligence (AGI). While companies like OpenAI and Anthropic are heavily invested in scaling up LLMs – the technology powering chatbots like ChatGPT – LeCun argues that this approach is fundamentally limited. He believes that genuine intelligence requires a deeper understanding of the world, not just the ability to predict the next word in a sequence. “The idea that you’re going to extend the capabilities of LLMs to the point that they’re going to have human-level intelligence is complete nonsense,” LeCun stated in an interview with Wired. This divergence in strategy has positioned AMI as a key player in an emerging alternative to the dominant LLM paradigm.
The $1.03 billion funding round was led by a consortium of investors including Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, the venture capital firm founded by Amazon’s Jeff Bezos. High-profile individuals also participated, including Mark Cuban, Xavier Niel, and former Google CEO Eric Schmidt, demonstrating broad support for LeCun’s approach. The company plans to establish a global presence with offices in Paris, Montreal, Singapore, and New York, with LeCun continuing his role as a professor at New York University alongside his leadership at AMI. This international footprint reflects a deliberate strategy to tap into diverse talent pools and foster a collaborative research environment.
Beyond Language: The Promise of World Models
The core of AMI’s strategy lies in the development of “world models.” These AI systems are designed to learn a comprehensive understanding of how the physical world operates – encompassing physics, object permanence, and causal relationships. Unlike LLMs, which excel at processing and generating text, world models aim to build an internal representation of reality, enabling them to reason about and interact with the environment in a more sophisticated way. This approach draws inspiration from how humans learn and understand the world, grounding intelligence in sensory experience and physical interaction.
LeCun envisions applications for these world models across a wide range of industries. He cites manufacturing, biomedical engineering, and robotics as key areas where AMI’s technology could have a transformative impact. For example, a world model could be used to simulate the operation of a complex machine like an aircraft engine, allowing engineers to optimize its performance, predict potential failures, and reduce emissions. The potential for these models extends beyond simulation, enabling robots to navigate complex environments, perform intricate tasks, and collaborate with humans more effectively. The company’s focus on practical applications is a key differentiator, aiming to translate cutting-edge research into tangible real-world solutions.
The development of world models isn’t entirely new. LeCun himself began exploring this area during his tenure at Meta, where he founded the Fundamental AI Research (FAIR) lab. Meta’s Joint-Embedding Predictive Architecture (JEPA) is an example of a world model developed within FAIR, demonstrating the feasibility of this approach. However, LeCun ultimately decided that pursuing this research outside of Meta would be more effective. He explained that Meta’s strategic shift towards LLMs, driven by competition in the rapidly evolving AI landscape, didn’t align with his long-term vision for world models. “There was a reorientation of Meta’s strategy where it had to basically catch up with the industry on LLMs and kind of do the same thing that other LLM companies are doing, which is not my interest,” LeCun told Wired.
A New Leadership Team
AMI has assembled a team of experienced AI researchers and entrepreneurs to lead its development. Alexandre LeBrun, formerly the CEO of the AI-powered healthcare startup Nabla, will serve as AMI’s CEO. Saining Xie, a former researcher at Google DeepMind, has been appointed as the company’s chief science officer. The founding team also includes Michael Rabbat, Meta’s former director of research science; Laurent Solly, former vice president of Europe at Meta; and Pascale Fung, former senior director of AI research at Meta. This combination of academic expertise, entrepreneurial experience, and industry knowledge positions AMI for rapid growth and innovation.
LeCun’s departure from Meta in November 2025, after 12 years with the company, marked a significant moment in the AI landscape. His decision to launch AMI represents a bold move to pursue his vision for the future of AI, independent of the pressures and priorities of a large technology corporation. While Meta plans to continue collaborating with AMI, LeCun believes that operating outside of the company will allow him to accelerate research and development, attract top talent, and forge partnerships with a wider range of organizations. The collaboration with Meta, as described by LeCun, will allow him to continue research while maintaining ties to his former employer.
The Debate Over AI’s Future
AMI’s approach directly challenges the prevailing belief among many AI labs, including OpenAI and Anthropic, that scaling up LLMs will ultimately lead to AGI. These companies are investing heavily in increasing the size and complexity of their language models, believing that this will unlock emergent capabilities and eventually result in human-level intelligence. However, LeCun argues that this is a misguided approach, akin to “building bigger and bigger statistical parrots.” He contends that LLMs, while impressive in their ability to generate text, lack the fundamental understanding of the world necessary for true intelligence.
The debate over the best path to AGI is likely to intensify as AMI begins to demonstrate the capabilities of its world models. The success of AMI could shift the focus of AI research away from LLMs and towards more embodied and grounded approaches. This could have profound implications for the development of AI-powered robots, autonomous systems, and other applications that require a deep understanding of the physical world. The company’s success will depend on its ability to translate its research into practical applications and demonstrate the superiority of its approach over competing technologies.
The funding secured by AMI provides a strong foundation for its ambitious goals. With a team of world-class researchers, a clear vision, and substantial financial backing, the company is well-positioned to become a leading force in the next generation of AI. The coming years will be crucial as AMI works to develop and deploy its world models, potentially reshaping the future of artificial intelligence. The company’s progress will be closely watched by the AI community and the broader technology industry.
AMI is currently focused on building its team and establishing its research infrastructure. The company plans to release more details about its specific projects and applications in the coming months. Investors and industry observers will be looking for concrete demonstrations of the capabilities of AMI’s world models and evidence that its approach can deliver on its promise of more intelligent and capable AI systems. The next major milestone for AMI will be the demonstration of a functional prototype showcasing the benefits of its world model technology.
Key Takeaways:
- Advanced Machine Intelligence (AMI) has raised $1.03 billion at a $3.5 billion valuation to develop AI “world models.”
- Founded by former Meta chief AI scientist Yann LeCun, AMI aims to create AI that understands the physical world, not just language.
- This approach challenges the current focus on large language models (LLMs) like ChatGPT.
- AMI plans to establish a global presence with offices in Paris, Montreal, Singapore, and New York.
- The company’s technology has potential applications in manufacturing, biomedical engineering, and robotics.
Stay tuned to World Today Journal for further updates on AMI’s progress and the evolving landscape of artificial intelligence. Share your thoughts on the future of AI in the comments below.