Tesla has halted its Dojo supercomputer project, which it had been developing for years to train autonomous driving systems. Company CEO Elon Musk stated on the social media platform X that scaling two different AI chips simultaneously was unwise. Musk’s announcement effectively shut down the Dojo project, which was based on Tesla’s custom chip design.

Dojo Project Ends

Dojo represented Tesla’s proprietary AI computing infrastructure developed to reduce its reliance on chip manufacturers like Nvidia and AMD. The project developed specialized hardware offering supercomputer-level processing power.

However, departures within the project’s technical leadership heralded long-standing uncertainties. Chip architect Jim Keller, who founded the Dojo project and previously worked at companies like Apple and AMD, left Tesla in 2020.

Ganesh Venkataramanan, who took over the project in his place, left in 2023. Team leader Peter Bannon is also leaving the company. It was reported that 20 engineers from the Dojo team have moved on to another AI company, DensityAI.

Elon Musk confirmed the project’s end and stated that the company is focusing on its own AI5 and AI6 chips. According to Musk, these chips will be used for both inference and model training. The company will continue to receive support from Nvidia and AMD to meet its high processing power needs.

So, what are your thoughts on this? Share your thoughts with us in the comments section below.