‘xAI Is a Failure’: AI Godfather Warns of Trouble for Musk’s AI Startup and a Bigger AI Bubble

Since its launch three years ago, Elon Musk has pitched xAI as a challenger to OpenAI, Anthropic, and Google in the race to build the most advanced artificial intelligence systems. One of the field’s most influential researchers is not convinced.

Yann LeCun, often referred to as one of the “godfathers of AI,” delivered a blunt assessment of Musk’s AI company during an interview with CNBC, calling xAI “kind of a failure” and questioning its ability to compete with the industry’s leading labs. His criticism went far beyond Musk’s startup. LeCun argued that many of the biggest AI companies are spending money at a pace that may prove difficult to sustain, raising the prospect of a broader shakeout across the industry.

LeCun’s comments add fresh fuel to a years-long public feud between the veteran AI researcher and Musk. The two have repeatedly clashed over the future of artificial intelligence, social media, and the direction of the tech industry.

“XAI is kind of a failure, frankly, because the founding team has departed, LeCun said.

He suggested Musk faces a growing challenge in attracting elite AI researchers.

“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.”

His remarks come after several xAI co-founders left the company over the past year. In February, Musk merged SpaceX and xAI in a deal that valued the combined business at $1.25 trillion.

Financial results released since that merger have also drawn attention. During the three months ended March 31, SpaceX’s AI segment, which includes xAI, reported an operating loss of $2.5 billion.

LeCun argued that xAI’s large-scale infrastructure investments have become a necessity rather than an advantage.

“xAI has huge infrastructure,” he said, adding that the company rents computing capacity to outside customers. “Because that’s the only way he [Musk] can recoup the cost.”

The comments refer to xAI’s Colossus 1 and Colossus 2 data centers in Memphis, Tennessee. The facilities house large clusters of AI chips and have attracted customers including Google and Anthropic, which have rented compute capacity.

Even with that infrastructure, LeCun remains skeptical about xAI’s long-term position.

“I’m not very positive about the prospect of xAI,” he said, adding that he does not expect the company to compete effectively with OpenAI or Anthropic on the frontier of AI development.

Neither SpaceX nor xAI immediately responded to CNBC’s request for comment.

A warning about AI economics

LeCun’s criticism extended beyond any single company. He questioned whether the economics supporting today’s AI boom can hold up as infrastructure costs continue to climb.

Enterprise spending on AI has drawn increased scrutiny in recent months. Businesses that rushed to adopt generative AI tools are now examining whether the benefits justify the growing costs.

OpenAI CEO Sam Altman recently acknowledged that AI expenses remain a major concern, saying during a company livestream that customers are actively discussing how much they spend on AI services.

LeCun believes the math is becoming difficult for many AI providers.

“The prices are going up of those AI services, but the cost of running them is going down, but not nearly fast enough. 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 very long, right?” he said.

His conclusion was stark.

“The labs like OpenAI and Anthropic are going to have to increase prices, they’re going to have to cut costs, or there’s going to be a big bubble explosion.”

The comments arrive at a time when investors have poured hundreds of billions of dollars into AI infrastructure, data centers, and model development. OpenAI, Anthropic, xAI, Google, Meta, and other major players continue to invest aggressively in pursuit of larger, more capable systems.

Why LeCun is betting on world models

LeCun has long argued that large language models, the technology behind ChatGPT, Claude, and Grok, are unlikely to deliver human-level intelligence on their own.

His preferred approach centers on what researchers call “world models.” Rather than learning patterns in text, world models attempt to build an internal representation of how the physical or simulated world works. The goal is to help AI systems reason about objects, actions, cause and effect, and real-world outcomes.

“I personally don’t think we’re going to have generalized reliable agentic systems until they’re based on world models,” LeCun said.

That view puts him at odds with many leading AI companies that are investing heavily in autonomous AI agents built primarily on large language models.

LeCun acknowledged that LLMs have proven highly effective for coding, mathematics, and other specialized tasks. His concern centers on cost.

“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.”

That tension between capability and profitability may become one of the defining questions facing the AI industry over the next few years. Investors have largely focused on who can build the most powerful models. LeCun is asking a different question: who can build them and still make money?

If his warning proves correct, the next chapter of the AI race may be shaped less by technological breakthroughs and more by economic reality.