A new artificial intelligence startup founded by prominent AI researcher Yann LeCun has raised more than $1 billion in funding as it works to develop a new class of intelligent systems designed to reason, plan, and better understand the real world.
The company, Advanced Machine Intelligence (AMI), announced that it has secured $1.03 billion in a fresh round of investment, giving the venture a pre-money valuation of $3.5 billion. The substantial funding reflects strong investor interest in emerging AI technologies that could move beyond the limitations of current systems built around large language models.
Several prominent investment firms participated in the funding round. The investment was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, the venture investment arm linked to Amazon founder Jeff Bezos.
The funding marks one of the largest investments in a newly formed AI startup and highlights growing interest in technologies that could push artificial intelligence closer to human-like reasoning and decision-making.
LeCun’s New Chapter After Leaving Meta
LeCun is widely regarded as one of the pioneers of modern artificial intelligence. Before launching AMI, he spent more than a decade at Meta Platforms, where he served as the company’s chief AI scientist.
He joined the company in 2013 when it was still known as Facebook and played a key role in building its internal AI research efforts. At that time, he founded Facebook AI Research, commonly referred to as FAIR, which later became one of the most influential research labs in the field.
Under his leadership, the lab produced significant breakthroughs in deep learning and machine learning technologies that are now widely used across the tech industry.
LeCun departed Meta at the end of 2025 and soon afterward began focusing on his new venture. With AMI, he is attempting to explore a different approach to artificial intelligence—one that focuses on reasoning and real-world understanding rather than relying primarily on language prediction models.
Moving Beyond Today’s AI Models
Over the past several years, the rapid development of large language models has transformed the AI landscape. These systems power a wide range of digital tools, including chatbots, content generators, and coding assistants.
However, LeCun has long argued that such models have fundamental limitations.
Most large language models operate by analyzing massive datasets and predicting the next word, image element, or piece of data in a sequence. While this method can produce impressive results, LeCun believes it does not necessarily lead to systems that truly understand the world around them.
AMI plans to pursue a different strategy.
The company is focused on building AI systems based on reasoning, planning, and so-called “world models.” These models attempt to simulate how the physical world functions, allowing machines to anticipate outcomes and make more informed decisions.
The idea is that AI should not simply replicate patterns found in data but should instead develop an internal understanding of how environments behave.
By doing so, the systems could potentially operate more effectively in complex real-world situations.
Targeting Industries With Complex Systems
In its early stages, AMI intends to focus on enterprise and industrial applications rather than consumer software.
The company believes its technology could be particularly useful for organizations that operate large and complicated systems requiring sophisticated decision-making.
Potential customers could include companies in sectors such as manufacturing, automotive production, aerospace engineering, biomedical research, and pharmaceuticals.
These industries often rely on intricate processes, automation systems, and data-driven planning tools. More advanced AI systems capable of reasoning about complex scenarios could help optimize operations, improve safety, and enhance productivity.
Rather than offering a single standalone product, AMI is expected to develop platforms that companies can integrate into their own systems. These tools could support a wide range of applications, from robotics and industrial automation to research analysis and supply-chain planning.
Long-Term Vision Includes Consumer Technology
Although AMI’s initial focus will be on enterprise use cases, LeCun believes the company’s research could eventually enable consumer-facing technologies as well.
One example involves the development of domestic robots capable of performing everyday tasks inside homes.
Creating such machines would require AI systems that can understand physical spaces, interact with objects, and anticipate the consequences of their actions. This type of common-sense reasoning remains one of the major challenges in artificial intelligence.
If AI systems can develop stronger models of the real world, they could eventually support a wide range of new consumer products.
Household robots, personal assistants that understand physical environments, and more advanced smart devices could all emerge from this type of technology.
Potential Applications in Wearable Devices
Despite leaving Meta, LeCun has indicated that he is still discussing potential collaborations with the company.
One area where AMI’s technology could potentially be used is in smart glasses and other wearable devices.
Meta has already introduced the Ray-Ban Meta smart glasses, created in partnership with the eyewear brand Ray-Ban. These devices combine cameras, microphones, and artificial intelligence features to help users capture photos and videos or interact with digital services.
Integrating AMI’s reasoning-focused AI technology into such hardware could allow wearable devices to better interpret the physical environment around users.
That could enable features such as more intelligent visual assistance, context-aware interactions, and improved real-world navigation.
While such applications remain speculative, they illustrate the broader range of possibilities that AMI’s technology could support.
