The car manufacturers efforts to generate unique system for driverless technology is about to come to an end since Tesla plans to dissolve the team behind its Dojo supercomputer. According to the claims of Bloomberg, which named unidentified sources, Peter Bannon of Dojo’s lead, is departing the firm, and the other team members will be moved to other data center and computing initiatives within Tesla.
The termination of Tesla’s Dojo initiatives comes after about 20 employees departed the carmaker to found DensityAI, an AI startup. In accordance with reports, soon the new business will come to light and will be manufacturing chips, hardware, along with the software that will power AI data centers used in robotics, by AI agents, as well as in automotive applications. Former Dojo CEO Ganesh Venkataramanan and former employes from Tesla Bill Chang and Ben Floering established DensityAI.
CEO of Tesla, Elon Musk has pushed for shareholders to recognize Tesla as an AI and robotics company, in spite of a limited robotaxi launch in Austin this past June that included Model Y cars with a human in the front seat of the car and led to multiple documented cases of the cars demonstrating hazardous driving behavior.
Since 2019, Musk has been having discussions about closing Dojo, so the decision of Tesla to do so shows a serious change in approach. Because Dojo can “process truly vast amounts of video data,” Musk has stated that it will be the cornerstone of Tesla’s AI goals and its objective to achieve complete self-driving.Even at the company’s second-quarter results call, he mentioned Dojo, albeit in passing.
Morgan Stanley anticipated in 2023 that Dojo might boost the market value of the company by $500 billion by generating fresh revenue streams which included software services along with robotaxis. Musk revealed last year that in the run-up to Tesla’s robotaxi unveiling in October, the company’s AI team would “double down” on Dojo.
On the other hand, Musk started to promote Cortex, Tesla’s “giant new AI training supercluster being built at Tesla HQ in Austin to solve real-world AI,” put a stop to discussions about Dojo in August 2024.
The Dojo project combined in-house chip manufacturing with a supercomputer. In 2021, at its first ever AI Day, Tesla formally introduced Dojo and showcased its D1 chip. Venkataramanan unveiled the chip, which Tesla stated would power the Dojo supercomputer when coupled with GPU of Nvidia. In addition, the car manufacturer disclosed that it had been working on creating a next-generation D2 microprocessor. The said microprocessor would correct anything wrong related to its predecessor’s information flow.
According to sources who spoke to Bloomberg, Tesla now intends to rely more on Nvidia and other outside tech partners, including as AMD for computing and Samsung for chip production. In an effort to create its AI6 inference chips, which were anticipated to scale from powering FSD and Optimus humanoid robots of Tesla to high-performance AI training in data centers, the company secured a $16.5 billion contract with Samsung last month.
Musk indicated about future reduction in staff during second-quarter earnings call of Tesla. Musk said:
“Thinking about Dojo 3 and the AI6 inference chip, it seems like intuitively, we want to try to find convergence there, where it’s basically the same chip”.
The announcement comes as the board of Tesla offers Musk a $29 billion compensation plan to stay at Tesla and support the company’s AI initiatives rather than being overly distracted by his other businesses, which include the more pure-play AI startup xAI.
