LeCun Slams Musk’s xAI, Warns of ‘Bubble Explosion’

Yann LeCun, a titan in the AI field and the founder of AMI Labs, has sharply criticized Elon Musk’s artificial intelligence venture, xAI, labeling it a “failure” that is ill-equipped to compete at the cutting edge of AI development. LeCun, formerly Meta’s chief AI scientist, articulated his concerns in a recent interview, warning of a potential “big bubble explosion” within the rapidly expanding AI industry.

These remarks further escalate a long-standing public disagreement between LeCun and Musk, casting a shadow over the formidable valuations commanded by some of the world’s leading AI enterprises.

LeCun, widely regarded as a “godfather of AI” for his pioneering work, has repeatedly clashed with Musk over the past few years. Their disputes have spanned topics from AI ethics to what LeCun has described as Musk’s “conspiracy theories” on social media. Musk, in turn, has accused LeCun of being “out of touch with AI for a long time.”

LeCun stated, “xAI is kind of a failure, frankly, because the founding team has departed.” He elaborated, “Elon is now in a position that is very, very difficult for him to kind of hire top people in AI, because he’s kind of, you know, not behaved in sort of very good ways toward the … previous team.”

Over the past year, xAI has seen several of its co-founders depart. In a significant move earlier this year, Musk merged SpaceX with xAI, a deal that reportedly valued the combined entity at a staggering $1.25 trillion. However, financial disclosures reveal that in the three months ending March 31, SpaceX’s AI segment, which encompasses xAI, incurred an operational loss of $2.5 billion. This contrasts with AMI Labs’ recent success, which raised $1 billion in a March funding round at a pre-money valuation of $3.5 billion, dedicated to the pursuit of world models.

LeCun highlighted that xAI possesses “huge infrastructure” which it leases to other companies, suggesting this is the primary mechanism for Musk to recoup costs. This refers to xAI’s Colossus 1 and Colossus 2 data centers in Memphis, facilities where both Google and Anthropic have secured compute capacity rentals.

“I’m not very positive about the prospect of xAI,” LeCun admitted, expressing skepticism about its ability to contend with industry giants like OpenAI and Anthropic.

Enterprise spending on artificial intelligence has increasingly come under scrutiny as the technology’s operational costs are proving to be substantially higher than initially anticipated. OpenAI CEO Sam Altman reportedly acknowledged this month in a company live stream that the escalating costs of AI services are a “huge issue” for many organizations.

“The prices are going up of those AI services, but the cost of running them is going down, but not nearly fast enough,” LeCun observed. “And so all of those companies are losing money, and basically, the use for most people is funded by the investors. That can’t go on for a very long right?”

The AMI Labs founder warned that leading AI labs such as OpenAI and Anthropic will face difficult choices. “They’re going to have to increase prices, they’re going to have to cut costs, or there’s going to be a big bubble explosion,” he stated.

LeCun has been a consistent critic of the limitations inherent in large language models (LLMs), the foundational technology powering many of today’s leading AI products. He advocates for an alternative paradigm: “world models.”

While LLMs excel at predicting the next word in a sequence, enabling sophisticated reasoning and coding capabilities, world models aim to build a deeper, causal understanding of how the real or simulated world operates, encompassing objects, cause-and-effect relationships, and actions. LeCun firmly believes, “I personally don’t think we’re going to have generalised reliable agentic systems until they’re based on world models.”

Currently, major AI players like Anthropic and OpenAI are investing heavily in AI agents—systems designed to autonomously perform complex tasks. However, LeCun points out a critical economic disconnect. He acknowledges the utility of LLMs for tasks like coding and mathematics, but emphasizes that “the cost of running those systems with this kind of performance is very high compared to the amount of money that users are ready to pay.” This economic reality underpins his concern about the sustainability of current AI business models and the potential for market correction.