Tesla dissolves Dojo team, shifts AI strategy amid talent exodus
Tesla is reportedly dismantling the team responsible for developing its Dojo supercomputer, a move signaling a significant strategic pivot away from internal technology for training the artificial intelligence models that power its self-driving features and Optimus humanoid robot. The dissolution of the team is said to coincide with the departure of Peter Bannon, who led the Dojo initiative.
Introduced by CEO Elon Musk in 2019, Dojo was envisioned as a “super powerful training computer” designed to process vast quantities of vehicle video data, essential for refining Tesla’s AI chips. However, recent reports suggest a shift in this vision, with Tesla now planning to leverage external partners for its computing needs, turning to Nvidia and AMD, while Samsung is slated to handle chip manufacturing.
This strategic redirection follows a notable exodus of talent from the Dojo unit earlier this month, when approximately 20 team members reportedly left Tesla to establish DensityAI, a new startup focused on data center services for sectors ranging from automotive to robotics. The remaining employees from the disbanded Dojo team have since been reassigned to other departments within the automaker.
The decision to scale back in-house AI hardware development comes on the heels of a major announcement last month, when Musk revealed a substantial $16.5 billion deal with Samsung for the development of Tesla’s next-generation AI6 chip. Responding to the news of Dojo’s dissolution, Musk affirmed on X that the “Tesla AI5, AI6 and subsequent chips will be excellent for inference and at least pretty good for training. All effort is focused on that.” This statement suggests a continued commitment to advanced AI chips, albeit with a refined approach to their development and manufacturing.
The disbandment of the Dojo team unfolds at a critical juncture for Tesla, particularly as the company expands its robotaxi service. While the recent launch in Austin, Texas, includes a safety monitor in the passenger seat, the subsequent rollout in San Francisco controversially places a human driver at the wheel. This latter approach appears to contradict Musk’s long-standing promise of a fully autonomous vehicle with “no one in the car,” underscoring the ongoing complexities and challenges in realizing true self-driving capabilities. The pivot in AI strategy, therefore, is not merely a technical adjustment but a potentially far-reaching decision with implications for Tesla’s core product offerings and its ambitious future vision.