Yann LeCun has never been quiet about his doubts, and his latest target is one of the most valuable companies on earth.
Elon Musk’s xAI is “kind of a failure” that will not be able to compete on the frontier with OpenAI and Anthropic, LeCun, the founder of AMI Labs and a former chief AI scientist at Meta, told CNBC.
His reasoning is about people. “The founding team has left, or was fired,” he said, conceding some uncertainty, and Musk now finds it “very, very difficult” to hire top AI talent “because he’s kind of, you know, not behaved in sort of very good ways” toward the previous team.
LeCun also took aim at xAI’s economics. The company has “huge infrastructure” it rents to others, he said, “because that’s the only way he can recoup the cost.”
The numbers behind the jab
He has a point on the infrastructure. xAI’s Colossus data centres in Memphis are now partly let out to rivals: Anthropic and Google both rent compute there, with Google alone paying SpaceX around $920mn a month.
The context is steep. In February, Musk merged SpaceX with xAI in a deal valuing the combined company at $1.25tn. In the first quarter, SpaceX’s AI segment, which includes xAI, posted a $2.5bn operating loss.
The bubble warning
LeCun’s bigger claim was about the whole industry. AI services are getting more expensive to run than the labs can charge for, he argued, and the gap is being papered over by investors.
“The prices are going up… 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,” he said. The labs, he added, “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.”
It is not a fringe worry. OpenAI’s Sam Altman recently called AI costs a “huge issue,” and enterprise spending on the technology has come under growing scrutiny.
The conflict worth naming
Here is the part to keep in mind. LeCun is not a neutral observer.
In March, his startup AMI Labs raised about $1bn on the thesis that the large language models OpenAI, Anthropic, and xAI all sell are a dead end. He argues the future belongs to “world models”, a different approach, and that “we’re not going to have generalized reliable agentic systems” without them.
So his warning of a bubble and his dismissal of xAI double as a pitch for his own alternative. It does not make him wrong, the unit-economics problem he describes is widely acknowledged, but it makes him the least disinterested person to say it. He and Musk have also feuded for years; Musk has called him “out of touch with AI for a long time.” xAI and SpaceX did not comment.
