According to Bloomberg, Tesla has broken up the team behind Dojo Supercomputer and has finished playing the automaker in developing in-house chips for driverless technology.
Dojo lead Peter Bannon leaves the company, and the remaining team members will be reassigned to other data centers to calculate projects within Tesla, according to a Bloomberg report that cited an anonymous source.
The dissolution of Tesla’s dojo efforts followed the eviction of around 20 workers, leaving the car manufacturer and establishing its own AI company called Dencesai. The new startup reportedly is coming out of Stealth soon, building chips, hardware and software that powers AI data centers used in robotics, AI agents and automotive applications. Desancy was founded by former dojo head Ganesh Venkataramanan and former Tesla employees Bill Chan and Ben Floating.
It also comes at a critical time for Tesla.
CEO Elon Musk pushed him to acquire shareholders to consider Tesla as an AI and Robotics Company this June despite limited launches of Robotakhi in Austin.
Tesla’s decision to close Dojo, which Musk has been talking about since 2019, is a major change in strategy. Musk says that “Dojo is the cornerstone of Tesla’s AI ambitions and that the goal is to reach fully automated driving to “process a truly enormous amount of video data.” He spoke about Dojo, albeit briefly, as well as the company’s second quarter revenue call.
In 2023, Morgan Stanley predicted that Dojo could add $500 billion to the company’s market value by unlocking new revenue streams in the form of Robotaxis and software services. Just last year, Musk said that Tesla’s AI team would “double” Dojo, led by Tesla’s Robotaxi Reveal, which took place in October.
TechCrunch Events
San Francisco
|
October 27th-29th, 2025
But we’ll talk about Dojo halting around August 2024, when Musk began to promote the cortex instead. Tesla’s “A giant new AI training super cluster built at Tesla’s headquarters in Austin solves real-world AI.”
The Dojo project was part of a supercomputer and part of the company’s chipmaking. Tesla unveiled its D1 chip when it officially unveiled Dojo on its first AI day of 2021. Venkataramananan has announced the chips. The automaker also said it is working on a next-generation D2 chip that solves the information flow bottleneck of its predecessor.
Sources told Bloomberg that Tesla is currently planning to increase its dependence on Nvidia, and that other external technology partners, such as Compute for Compute AMD and Samsung, will also rely on Chip Manufacturing. Tesla signed a $16.5 billion deal with Samsung last month to create an AI6 inference chip. This is a chip design that promises scaling from expanding the power of FSD and Tesla’s Optimus Humanoid robots to high-performance AI training in the data center.
During Tesla’s second quarter revenue calls, Musk suggested potential redundancy.
“When you think about Dojo 3 and AI6 inference chips, intuitively, you try to find convergence, which is basically the same chip,” says Musk.
The news is as Tesla’s board is offering a $29 billion salary package to help Tesla keep him on mask and move the company’s AI efforts forward rather than falling too aside to his other companies, including purely play AI startup Xai.
TechCrunch contacts Tesla for more details.
Do you have sensitive tips or confidential documents? We report on the internal mechanisms of the AI industry. From companies shaping their futures to those affected by their decisions. Contact Rebecca Bellan and Maxwell Zeff at maxwell.zeff@techcrunch.com. For secure communication, please contact us via the signals @rebeccabellan.491 and @mzeff.88.
