Key Points
- Tesla is disbanding the Dojo supercomputer team, with lead Peter Bannon exiting.
- About 20 former Dojo employees left to start the stealth AI startup DensityAI.
- Remaining Dojo staff will be reassigned to other data‑center and compute projects.
- Tesla is shifting toward external partners like Nvidia, AMD and Samsung for AI hardware.
- A multi‑billion‑dollar deal with Samsung will produce AI6 inference chips for various applications.
- The change marks a major strategic pivot away from in‑house chip development for FSD and robotaxis.
Background
Since its announcement in 2021, Tesla’s Dojo supercomputer has been positioned as a cornerstone of the company’s AI ambitions, promising to process vast amounts of video data to accelerate full self‑driving (FSD) development. The project combined a custom supercomputer architecture with in‑house chip design, highlighted by the D1 chip unveiled at Tesla’s first AI Day. Over the years, Tesla has highlighted Dojo’s potential to add significant market value by unlocking new revenue streams from robotaxis and software services.
Team Disbandment
According to Bloomberg, Tesla is dismantling the Dojo team. Lead Peter Bannon is leaving the company, and the remaining members will be reassigned to other data‑center and compute initiatives within Tesla. The restructuring follows the departure of roughly 20 employees who left to form the stealth AI startup DensityAI. DensityAI, founded by former Dojo head Ganesh Venkataramanan and ex‑Tesla engineers Bill Chang and Ben Floering, aims to build chips, hardware and software for AI data centers used in robotics, AI agents and automotive applications.
Implications for Tesla’s AI Strategy
The decision signals a shift away from Tesla’s in‑house chip development toward greater reliance on external technology partners. Sources indicate Tesla will increase its dependence on Nvidia and AMD for compute, and Samsung for chip manufacturing. Recently, Tesla signed a multi‑billion‑dollar agreement with Samsung to produce AI6 inference chips, which are described as scalable from powering FSD and Optimus humanoid robots to high‑performance AI training in data centers. Elon Musk hinted at potential redundancies during the second‑quarter earnings call, suggesting a convergence between future Dojo iterations and the AI6 inference chip.
Industry Reaction
The move arrives as Tesla’s board offers Musk a substantial pay package to keep him focused on the company’s AI initiatives, while he also leads other ventures such as xAI. Analysts had previously projected that Dojo could add hundreds of billions to Tesla’s market value. The disbanding of the Dojo effort therefore represents a major strategic pivot, with investors and industry observers watching how Tesla will pursue its robotaxi and broader AI goals without the dedicated in‑house supercomputer team.
Future Outlook
While Dojo’s development is being halted, Tesla continues to pursue AI advancements through external collaborations and its new AI6 chips. The company’s focus appears to be on integrating these technologies into its existing vehicle platforms and future robotics projects. The long‑term impact of this restructuring on Tesla’s ability to achieve full self‑driving autonomy and launch a scalable robotaxi service remains to be seen.
Source: techcrunch.com