seniorspectrumnewspaper – Tesla has officially ended its Dojo supercomputer project, marking a significant change in its AI development strategy. As first reported by Bloomberg, the company is disbanding the team behind Dojo and reallocating remaining staff to other departments. This move signals Tesla’s decision to abandon its ambitious plan to build a proprietary AI training infrastructure in favor of more scalable chip solutions. Peter Bannon, who led the Dojo project, is also leaving the company. His exit follows the departure of around 20 other Dojo team members who recently founded a new startup, DensityAI.
Read More : Google Fixing Gemini Bug That Triggers Self-Criticism
Tesla originally launched Dojo to train AI models powering autonomous driving and its humanoid robot project, Optimus. However, the company has now decided to discontinue the effort and consolidate its resources into developing next-generation chips: AI5 and AI6.
In response to the news on X, Tesla CEO Elon Musk explained the decision. Stating, “It doesn’t make sense for Tesla to divide its resources and scale two quite different AI chip designs.” Musk emphasized that Tesla’s in-house chips will serve well for inference and will be adequate for training.
The shutdown of Dojo ends a multi-year effort to build a custom AI training architecture. Which Tesla had once touted as a breakthrough for real-time machine learning and vision systems. Instead, the company will increase reliance on external vendors like NVIDIA for training-specific hardware. Spending billions to support its AI infrastructure.
AI5 and AI6 Chips to Power Tesla’s Autonomous Systems and Robotics
With Dojo shelved, Tesla is concentrating on its AI5 and AI6 platforms to drive the future of its autonomous vehicles and robotics. The AI6 chip, developed in collaboration with Samsung under a $16.5 billion deal, is expected to play a central role. These chips will power real-time inference capabilities onboard Tesla cars and Optimus robots.
Musk noted that while these chips are optimized for inference, they will still be “pretty good” for training purposes. AI5 is scheduled to enter production in 2026, followed by AI6, which will push performance even further. This pivot aligns Tesla more closely with other tech companies that use a hybrid approach—developing in-house hardware for specific needs while sourcing high-performance training silicon externally. By consolidating its AI chip efforts, Tesla aims to streamline development and scale more efficiently.
Read More : Uber Logged 400,000 Misconduct Reports Since 2017
Meanwhile, the company continues to experiment with AI in other domains. xAI’s Grok chatbot now available in Tesla vehicles. Is part of a broader push to integrate conversational AI into the driving experience. Tesla is also piloting its Robotaxi fleet, although early testing has produced mixed results.
The end of Dojo represents more than just a project closure—it reflects Tesla’s evolving approach to AI development, balancing ambition with practicality. As the company shifts focus to AI5 and AI6, its strategy suggests a more targeted, production-ready vision for deploying AI at scale across its automotive and robotics platforms.
