AMI Labs, a new artificial intelligence research company co-founded by Turing Award winner Yann LeCun, has raised $1.03 billion in funding at a $3.5 billion pre-money valuation.
The company is focused on developing “world models,” a new class of AI systems designed to learn directly from real-world data rather than relying primarily on language-based training. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with participation from multiple institutional and individual investors.
What The Company Does
AMI Labs is developing AI systems intended to build structured models of the physical and social world. These systems aim to understand how environments, objects, and events interact over time rather than simply predicting text.
The company’s approach builds on the Joint Embedding Predictive Architecture (JEPA), a research framework proposed by Yann LeCun in 2022. The architecture is designed to allow AI models to learn representations of the world by predicting how different parts of an environment relate to one another.
Unlike many current generative AI models that rely heavily on large-scale language training, world models are intended to incorporate visual, sensory, and environmental information to develop a deeper understanding of real-world dynamics.
AMI Labs’ research could eventually support applications across fields such as robotics, healthcare, and scientific modeling, where AI systems need to reason about physical environments rather than only textual information.
Advertisement
Market Context / Industry Background
The rapid growth of generative AI has been driven largely by large language models capable of generating text, images, and other forms of content. However, researchers have also highlighted limitations in these systems, including their tendency to produce inaccurate information, often referred to as hallucinations.
For certain applications—particularly in sectors such as healthcare—these limitations can present significant challenges. As a result, some researchers are exploring alternative architectures that may allow AI systems to develop more reliable representations of the world.
World models represent one such research direction. The concept focuses on enabling AI systems to learn predictive representations of real-world environments by combining different types of data, including images, sensor inputs, and contextual information.
While the category currently includes relatively few companies, investor interest has begun to grow. Several startups exploring world model architectures have recently raised large funding rounds, suggesting increasing attention from both research institutions and venture investors.
Founder / Investor Commentary
AMI Labs CEO Alexandre LeBrun said the company’s goal is to develop AI systems capable of understanding the world in a more comprehensive way than current language-based models.
“My prediction is that ‘world models’ will be the next buzzword,” LeBrun said. “In six months, every company will call itself a world model to raise funding.”
He emphasized that the company’s research agenda differs from many applied AI startups that aim to release products quickly.
“AMI Labs is a very ambitious project because it starts with fundamental research,” LeBrun said. “It’s not your typical applied AI startup that can release a product in three months, have revenue in six months and make $10 million in annual recurring revenue in 12 months.”
According to LeBrun, developing commercially viable world models may take several years as the underlying research advances.
Growth Plans / Use Of Funds
AMI Labs plans to use the funding primarily to support two major areas: computing infrastructure and recruitment of specialized AI researchers.
The company intends to build research teams across four key locations: Paris, where its headquarters are located; New York, where LeCun teaches at New York University; Montreal, where senior researcher Michael Rabbat is based; and Singapore, which the company views as a strategic hub for recruiting AI talent and building relationships with potential clients in Asia.
Although the company does not plan to generate revenue in the near term, it expects to collaborate with industry partners as its research progresses. One of its first disclosed partners is digital health company Nabla, where LeBrun serves as chairman.
The partnership may allow early testing of the company’s AI models in healthcare settings where reliable reasoning about real-world conditions is particularly important.
AMI Labs also plans to maintain a research-oriented approach by publishing academic papers and releasing portions of its code as open source.
“We will also make a lot of code open source,” LeBrun said. “We think things move faster when they’re open, and it’s in our best interest to build a community and a research ecosystem around us.”
About AMI Labs
AMI Labs is an artificial intelligence research company focused on developing world models capable of learning from real-world data. Founded by Yann LeCun and Alexandre LeBrun, the company conducts fundamental research into AI architectures designed to understand environments, physical systems, and complex interactions. Headquartered in Paris with research teams across multiple global locations, AMI Labs aims to build AI systems that go beyond language-based models to support applications in fields such as healthcare, robotics, and scientific research.
