Former Meta AI leader raises $1 billion to challenge the chatbot boom

A prominent artificial intelligence researcher is launching a bold challenge to the technology industry’s current obsession with chatbots.

According to an American English language website Quartz, Yann LeCun, who previously led AI research at Meta Platforms, has raised about $1.03 billion for a new startup called Advanced Machine Intelligence (AMI). His goal is to prove that the future of AI may lie somewhere very different from today’s popular chatbot systems.

For the past few years, most major tech companies have focused heavily on large language models (LLMs) – the technology behind chatbots that can answer questions and generate text. But LeCun has long argued that these systems, while impressive, don’t truly understand the world. Instead, they mainly predict the next word in a sentence based on patterns in data.

Now he wants to build something different.

A DIFFERENT PATH TO SMARTER AI

AMI aims to develop what researchers call “world models.” These are AI systems designed to understand how the real world works, including physical spaces, cause and effect, and how events unfold over time, rather than simply generating convincing text.

LeCun believes this approach could lead to machines that can reason, plan, and make decisions more reliably. His research framework for this idea is known as JEPA (Joint Embedding Predictive Architecture), which focuses on helping AI learn patterns and relationships in the real world.

Such technology could be especially useful in areas where mistakes can be costly, such as manufacturing, robotics, healthcare, and biomedical research. In those environments, producing a plausible but incorrect answer – something chatbots are known to do – could create serious problems.

LEAVING META TO PURSUE THE IDEA

LeCun spent more than a decade building the AI research division at Meta Platforms before leaving in late 2025. During that time, many tech companies were racing to build larger and more powerful chatbots.

According to LeCun, that race wasn’t the direction he wanted to pursue. He has said he believed he could develop his ideas “faster, cheaper, and better” outside the company, which led him to launch AMI.

The startup’s chief executive is Alexandre LeBrun. He says the company isn’t planning to release products quickly like many AI startups that rush to market.

Instead, the team is focused on building deeper technology first – even if it takes years.

A GROWING TREND IN AI RESEARCH

AMI isn’t the only group exploring this direction. In recent months, several researchers have begun promoting AI systems that understand the physical world rather than just text. For example, computer scientist Fei‑Fei Li recently raised about $1 billion for World Labs, a startup working on “spatial intelligence,” which helps AI interpret and create three-dimensional environments.

Major technology firms are also investing in the concept. Nvidia has described world models as a key tool for training robots and autonomous vehicles by simulating real-world situations.

BIG INVESTORS PLACE A BET

AMI’s funding round attracted support from a long list of major investors, including Nvidia, Samsung, Toyota Ventures, and Temasek Holdings. Individual backers include tech figures such as Mark Cuban and former Google CEO Eric Schmidt.

The company is headquartered in Paris, with plans for offices in New York, Montreal, and Singapore.

A RISKY BUT AMBITIOUS BET

Despite the huge investment, AMI currently has no product on the market and doesn’t expect quick results. Its strategy relies on the idea that today’s chatbot-focused approach to AI may not lead to truly intelligent machines.

If LeCun is right, the next major wave of artificial intelligence may come from systems that understand the world itself, not just the words people type into a prompt.

Whether that vision proves revolutionary – or simply very expensive – is something the tech industry will be watching closely.