June 22, 2026 Artificial intelligence pioneer Yann LeCun says Elon Musk’s xAI is unlikely to compete at the leading edge of the AI industry. LeCun also warned that rising costs and continued investor subsidies across the sector could eventually lead to a “big bubble explosion.”
LeCun, founder of AMI Labs and former chief AI scientist at Meta, made the remarks in an interview with CNBC, renewing a long-running public disagreement with Musk over artificial intelligence. “XAI is kind of a failure, frankly, because the founding team has departed,” LeCun said.
He added that Musk is in a difficult position when it comes to recruiting top AI researchers because of what LeCun described as the entrepreneur’s treatment of former team members.
Several xAI co-founders have left the company over the past year. In February, Musk merged SpaceX with xAI in a deal valuing the combined company at US$1.25 trillion. According to financial results for the three months ended March 31, SpaceX’s AI segment, which includes xAI, posted an operating loss of US$2.5 billion.
LeCun said xAI’s large infrastructure footprint is being rented to other companies as a way to offset costs. xAI operates the Colossus 1 and Colossus 2 data centres in Memphis, Tenn., with Google and Anthropic among companies that have rented computing capacity at the facilities.
“I’m not very positive about the prospect of xAI,” LeCun said, adding that he does not expect the company to compete successfully with OpenAI or Anthropic.
The godfather of AI also raised concerns about the broader economics of the AI industry, arguing that many companies continue to lose money despite charging users for services. “The prices are going up of those AI services, but the cost of running them is going down, but not nearly fast enough,” he said.
According to him, companies such as OpenAI and Anthropic will eventually need to raise prices, cut costs or face what he described as a significant market correction.
LeCun has long criticized large language models, or LLMs, which underpin products such as ChatGPT and Claude. Instead, he advocates for so-called “world models,” which aim to understand cause and effect and how the real world operates. “I personally don’t think we’re going to have generalized reliable agentic systems until they’re based on world models,” he said.
Many leading AI companies, including OpenAI and Anthropic, are currently focusing on AI agents capable of performing complex tasks autonomously. While LeCun said LLMs remain useful for applications such as coding and mathematics, he questioned whether their high operating costs can be sustained over the long term.
