The founder of AMI Lab, Yann LeCun, has labeled Elon Musk's xAI a "failed attempt," stating it will struggle to compete with giants like OpenAI and Anthropic.

The former chief AI scientist at Meta Platforms, Inc. warns that unless AI labs cut costs and raise prices, the sector faces a significant bubble burst.

LeCun's AMI Lab recently secured $10 billion in funding to develop world models, which he views as the core technology for AI's next evolution.

In an interview, AMI Lab founder Yann LeCun stated that Elon Musk's xAI is fundamentally a failed venture, unlikely to succeed in the competitive AI landscape. He also shared his perspective on the looming risk of a major industry bubble.

These comments escalate the long-running public feud between LeCun and Musk, raising market questions about the valuations of leading AI companies.

LeCun, the former chief AI scientist at Meta Platforms, Inc., has clashed with Musk for years over technical approaches and what he calls the Tesla Motors CEO's promotion of conspiracy theories on social media. Musk has countered that LeCun is out of touch with current industry developments.

LeCun is often called the "Godfather of AI" due to his foundational contributions to the field's early development.

"Frankly, xAI is a failure, primarily because the founding team members have left one after another," LeCun said.

"It's now extremely difficult for Musk to attract top AI talent because of how he treated the original founding team."

Over the past year, several xAI co-founders have departed. In February, Musk completed a major merger of SpaceX and xAI, valuing xAI at $1.25 trillion.

In the quarter ending March 31, the AI segment of SpaceX, which includes xAI, reported an operating loss of $2.5 billion. Meanwhile, LeCun's AMI Lab raised $10 billion in March, focusing on world model development with a pre-money valuation of $35 billion.

LeCun noted that xAI has built massive computing infrastructure and is now forced to recoup costs by leasing this capacity to other firms.

This infrastructure refers to xAI's "Colossus 1" and "Colossus 2" data centers in Memphis, with clients like Google and Anthropic leasing compute resources.

"I am not optimistic about xAI's prospects," LeCun said, believing it will be hard for the company to catch up to leaders OpenAI and Anthropic.

Industry Faces Imminent Bubble Burst

Corporate investment in AI is under intense scrutiny, as deployment costs have far exceeded initial industry expectations. Reports indicate OpenAI's CEO Sam Altman recently stated in an internal broadcast that companies are now carefully calculating AI investments, with high compute costs becoming a major challenge.

"While AI service pricing is rising, operational costs are not falling fast enough. Most AI companies are still losing money; user payments essentially come from investor funding, a model that is unsustainable in the long run," LeCun analyzed.

The AMI Lab founder warned that leading AI firms like OpenAI and Anthropic must either raise prices or cut costs, or the entire industry will eventually face a massive bubble burst.

LeCun has been a vocal critic of the limitations of large language models (LLMs), the foundation of current mainstream AI products, advocating instead for a new technical path: world models.

LLMs predict text by learning language patterns, excelling at logical reasoning and coding tasks. World models take a different approach, aiming to enable AI to understand the logic of the real or virtual world, learning object relationships, causality, and behavioral logic.

"Until technology based on world models is realized, it will be very difficult to develop generally capable, reliable autonomous agent systems," LeCun stated.

Currently, companies from Anthropic to OpenAI are heavily investing in developing agents—systems that can autonomously perform complex tasks.

LeCun acknowledges the utility of LLMs in scenarios like programming and mathematics but points out a fundamental business model flaw: the cost to maintain current model performance is prohibitively high, far exceeding what users are willing to pay.