Tesla shuts down Dojo, the AI training supercomputer that Musk said would be key to full self-driving

Tesla breaks the team behind his Dojo-Supercomputer and ends the game of the automaker in developing internal chips for technology without director, according to Bloomberg.

The lead of Dojo, Peter Bannon, leaves the company, and the remaining team members will be re -assigned to another data center and projects within Tesla, according to Bloomberg’s report, which quoted anonymous sources.

The dissolution of Tesla’s Dojo efforts follows the departure of around 20 employees, who have left the automaker to start their own AI Company called Densityai. The new startup is said to be from Stealth and builds chips, hardware and software that will provide data centers for AI with electricity that are used in robotics, by AI agents and in car applications. Densityai was founded by former Dojo head Ganesh Venkataramanan and former Tesla employees Bill Chang and Ben Floing.

It also comes at a crucial time for Tesla.

CEO Elon Musk has urged shareholders to show Tesla as an AI and Robotics company, despite a limited robotaxi launch in Austin last June with model Y vehicles with a human in the front passenger chair and resulted in a number of reported incidents of the vehicles that showed a problematic driving behavior.

Tesla’s decision to close Dojo, which Musk has talked about since 2019, is an important shift in the strategy. Musk said that Dojo would be the cornerstone of Tesla’s AI ambitions and his goal to achieve completely self-driving because of the ability to “really handle huge amounts of video data”. He spoke about Dojo, albeit briefly, as recently as the profit call of the company in the second quarter.

In 2023 Morgan Stanley predicted that Dojo could add $ 500 billion to the market value of the company by unlocking new income flows in the form of robotaxis and software services. Last year, Musk noted that the AI team of Tesla would “double” on Dojo in the run-up to the Robotaxi disclosure of Tesla, which took place in October.

But talking about Dojo stopped around August 2024, when Musk began to praise Cortex instead, Tesla’s “Giant New Ai Training Supercluster was built on Tesla HQ in Austin to solve Real-World AI.”

The DOJO project was a supercomputer part, a part in-house chip-making. Tesla unveiled his D1 chip when the Formal Dojo announced on the first AI day in 2021. Venkataramanan presented the chip, of which Tesla said it would be used in addition to Nvidia’s GPU to provide the Dojo -Supercomputer with power. The automaker also said it worked on a subsequent generation D2 chip that would solve any bottlenecks of the information flow from its predecessor.

Sources told Bloomberg that Tesla is now planning to increase its dependence on Nvidia, as well as other external technical partners such as AMD for Compute and Samsung for chip production. Tesla drew a $ 16.5 billion deal last month with Samsung to make its AI6 inference chips, a chip design that promises to scale from FSD and Tesla’s Optimus Humanoid robots all the way to high-quality AI training in data elphans.

During Tesla’s call for profit of the second quarter, Musk hinted at potential fired.

“Thinking of Dojo 3 and the AI6 Inference Chip, it seems intuitive, we want to try to find convergence where it is in fact the same chip,” said Musk.

The news is because the Tesla Musk board offers a payment package of $ 29 billion to keep him at Tesla and to help the AI efforts of the company move forward, instead of becoming sideways by its other companies, including the more pure AI-Startup Xai.

WAN has reached Tesla for more information.

Do you have a sensitive tip or confidential documents? We report on the inner operation of the AI industry – of the companies that shape its future to the people who are affected by their decisions. Please contact Rebecca Bellan on rebecca.bellan@techcrunch.com and Maxwell Zeff on maxwell.zeff@techcrunch.com. For safe communication you can contact us via Signaal on @Rebeccabellan.491 and @Mzeff.88.