Artificial intelligence systems built on large language models have advanced rapidly, but many researchers believe they still struggle with real-world reasoning and decision-making. A new startup founded by one of the most influential figures in AI is betting on a different path.
Advanced Machine Intelligence (AMI), the startup launched by former Meta chief AI scientist Yann LeCun, has raised $1.03 billion in funding at a $3.5 billion pre-money valuation, according to a Reuters report.
The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital and Bezos Expeditions.
The investment marks one of the largest funding rounds for a new AI research company this year.
The new capital will support research and development as well as efforts to turn the technology into deployable systems for industries that depend on reliable AI decision-making.
A challenge to the current AI model approach
AMI was founded by Yann LeCun, who previously served as Meta’s chief AI scientist and helped establish Facebook AI Research (FAIR). He left the company in late 2025.
The startup is built around LeCun’s long-held view that today’s large language models (LLMs) cannot by themselves achieve human-level intelligence.
Speaking to Reuters, LeCun said current systems that predict the next word or pixel are limited in their ability to understand the real world.
“Our main goal is to build intelligent systems that understand the real world,” LeCun told Reuters.
AMI is working on a different architecture called “world models,” which aims to help AI systems build abstract representations of real-world data and predict how environments evolve.
The idea is that these systems could reason about actions and plan tasks more reliably than current generative AI approaches.
Rather than targeting consumer chatbots initially, AMI plans to focus on organisations that manage complex physical systems.
These include manufacturers, automakers, aerospace companies, biomedical firms, and pharmaceutical companies.
According to Reuters, the company wants to develop AI tools capable of reasoning and planning in complex environments where safety, reliability and control are critical.
AMI believes such systems could be useful for industrial automation, robotics, wearable devices, healthcare technologies and other sectors that rely on sensor-based data.
