Elon Musk announced over the long weekend that Tesla intends to restart work on Dojo 3, the company’s previously abandoned third-generation AI chip. This time, however, the project’s goal has changed dramatically. Instead of training self-driving models on Earth, Musk stated Dojo 3 will be dedicated to “space-based AI compute.”
This revival comes just five months after Tesla effectively shut down its original Dojo effort. The company disbanded the team behind its Dojo supercomputer following the departure of Dojo lead Peter Bannon. Around twenty Dojo workers also left to join a new AI infrastructure startup called DensityAI, founded by former Dojo head Ganesh Venkataramanan and other ex-Tesla employees.
At the time of the shutdown, reports indicated Tesla planned to increase its reliance on partners like Nvidia and AMD for computing power and Samsung for chip manufacturing, moving away from developing its own custom silicon. Musk’s latest comments suggest the strategy has shifted once again.
Musk explained in a post on X that the decision to revive Dojo was based on the state of Tesla’s in-house chip roadmap. He noted that the company’s AI5 chip design was “in good shape.” The AI5 chip, manufactured by TSMC, was designed to power Tesla’s automated driving features and its Optimus humanoid robots. Furthermore, Tesla signed a major deal with Samsung last summer to build its next-generation AI6 chips.
Musk specifically said, “AI7/Dojo3 will be for space-based AI compute,” positioning the resurrected project as an ambitious moonshot. To achieve this, Tesla is now preparing to rebuild the team it dismantled months ago. Musk used his post to recruit engineers directly, inviting them to email the company with details of tough technical problems they have solved.
The timing of this announcement is notable. It follows Nvidia’s unveiling of an open-source AI model for autonomous driving at CES 2026, a product that directly challenges Tesla’s own Full Self-Driving software. Musk commented that solving the rare edge cases in driving is “super hard,” and expressed hope for Nvidia’s success.
Musk and several other AI executives have argued that the future of data centers may lie off-planet, as Earth’s power grids are increasingly strained. Reports indicate that even Musk’s rival, OpenAI CEO Sam Altman, is excited by the prospect of orbital data centers. Musk holds a potential advantage in this area because he controls the launch vehicles through SpaceX.
Reports suggest Musk plans to use the upcoming SpaceX IPO to help finance a vision of using Starship rockets to launch a constellation of computing satellites. These satellites would operate in constant sunlight, harvesting solar power continuously. However, significant roadblocks remain, such as the challenge of cooling high-power computing hardware in the vacuum of space.
Musk’s comments about building “space-based AI compute” fit a familiar pattern for the entrepreneur: floating an ambitious, far-fetched idea and then attempting to brute-force it into reality.