Tesla CEO Elon Musk has announced a strategic shift in the company’s AI chip development, confirming that Tesla will streamline its efforts to focus on inference chips.
According to reports, the chips are used to run AI models and make real-time decisions, and the decision is to discontinue work on two separate chip architectures.
The move follows a Bloomberg report stating that Musk ordered the disbanding of Tesla’s in-house Dojo supercomputer team, with team leader Peter Bannon reportedly departing the company. Tesla has not responded to requests for comment.
Dojo was originally designed to train Tesla’s autonomous driving software using massive video and data inputs from its electric vehicles.
However, Musk said in a post on X (formerly Twitter), “It doesn’t make sense for Tesla to divide its resources and scale two quite different AI chip designs.”
He added that future efforts will center on Tesla’s AI5 and AI6 chips, which he claims will be “excellent for inference and at least pretty good for training.”
The AI5 chips are expected to enter production by late 2026, while Musk recently announced a $16.5 billion deal with Samsung Electronics to supply AI6 chips, though no timeline was provided.
Morgan Stanley analysts previously valued the Dojo project at $500 billion, suggesting it could open a new revenue stream for Tesla beyond vehicle sales. It remains unclear how the restructuring will affect that valuation.
Tesla has faced significant internal changes over the past year, including executive departures, job cuts, and declining EV sales amid rising competition and consumer backlash in Europe.
The company is now doubling down on AI-driven technologies, including self-driving vehicles and its Optimus humanoid robots.
According to Bloomberg, around 20 members of the Dojo team have left to join a new venture, DensityAI, while remaining staff are being reassigned to other Tesla data center and compute initiatives.
