Tesla Shuts Down Dojo Supercomputer, Shifts AI Strategy to External Partners
Tesla has officially shut down its Dojo supercomputer, which was used to train the AI behind Autopilot, Full Self-Driving (FSD), and the Optimus robot. The decision, first reported by Bloomberg and later confirmed by CEO Elon Musk, marks a strategic shift as Tesla moves away from building in-house AI systems and instead partners with external chipmakers like Nvidia, AMD, and Samsung.
As part of this change, Peter Bannon, head of the Dojo project, is leaving Tesla. Other team members will be reassigned to different compute and data center roles within the company. Musk had hinted at this move during a recent earnings call, suggesting a focus on more unified AI chip architecture.
Despite scrapping Dojo, Tesla remains committed to its AI goals. The company is building a new, more powerful AI cluster called Cortex in Austin, Texas, powered by over 100,000 Nvidia H100 and H200 chips. Another data center is also active in Memphis.
Additionally, Tesla signed a $16.5 billion deal with Samsung to produce AI6 (Hardware 6) chips locally in Texas, securing long-term access to key AI semiconductors through 2033. The company is also working with TSMC on AI5 chips, solidifying its commitment to future autonomous driving technology.
As part of this change, Peter Bannon, head of the Dojo project, is leaving Tesla. Other team members will be reassigned to different compute and data center roles within the company. Musk had hinted at this move during a recent earnings call, suggesting a focus on more unified AI chip architecture.
Despite scrapping Dojo, Tesla remains committed to its AI goals. The company is building a new, more powerful AI cluster called Cortex in Austin, Texas, powered by over 100,000 Nvidia H100 and H200 chips. Another data center is also active in Memphis.
Additionally, Tesla signed a $16.5 billion deal with Samsung to produce AI6 (Hardware 6) chips locally in Texas, securing long-term access to key AI semiconductors through 2033. The company is also working with TSMC on AI5 chips, solidifying its commitment to future autonomous driving technology.
