‘AI Godfather’ LeCun: Musk’s xAI is a failure

“To be honest, xAI is a kind of failure, because the company’s founding team has left,” he commented in an interview with CNBC June 18. “Elon is in a very difficult position to attract top talent in the AI ​​field, because he has not treated the team in a really good way in the past.”

LeCun’s statement reignited the long-standing disagreement between him and the richest man on the planet. Yann LeCun once went on social network

In February, Musk decided to merge xAI into SpaceX – an aerospace technology company that currently has a capitalization of $2,400 billion. However, the first quarter 2026 financial report shows that SpaceX’s AI segment is suffering an operating loss of $2.5 billion.

“I’m not optimistic about xAI’s prospects. They have no chance to compete with OpenAI or Anthropic,” said Meta’s former AI scientific director and current founder of AMI Labs.

LeCun also warned of the “AI bubble burst” of the entire industry when operating costs are beyond the tolerance of businesses. Sam Altman, CEO of OpenAI, last week also admitted that costs are a “painful issue”.

“Output service prices have increased, operating costs have decreased, but the speed is too slow. Most AI companies are losing money and living off investors’ pockets. This situation cannot last long. If large laboratories do not find a way to cut costs, the AI ​​bubble will definitely burst,” he said.

Besides the economic problem, the famous scientist also mentioned the limitations of large language models (LLM), saying that they are strong in programming or solving problems thanks to the ability to predict words based on probability, but the maintenance fee is too expensive compared to the user’s level of payment. He advocates the World Model, aiming to help computers understand how the real world operates through the rules of cause – effect and actions, different from the current LLM.

“We will not be able to have a reliable, autonomous AI agent system if we stick to the LLM without moving on to building on world models,” he emphasized.

LeCun, 66 years old, is a famous scientist in the field of AI. He, along with professors Geoffrey Hinton and Yoshua Bengio, is known as the Godfather of AI. He joined Meta in 2013, heads FAIR – the team that created Meta’s first open source model called Llama – and is the company’s chief AI scientist. However, he left Mark Zuckerberg’s company in 2025 to create his own startup specializing in “world simulation capabilities”.

Before leaving Meta, LeCun assessed that LLM was useful but did not have the potential to be as intelligent as humans, and that the field of artificial intelligence was overhyped, especially with the race to develop LLM. To enter the next revolution, companies need to “take a step back”, find out what is missing in current approaches if they want to create human-level intelligent AI.