Article Contents:
- Introduction: A new focus after halting Dojo
- Musk’s announcement: An “epic” chip on the way
- Why was the Dojo project halted?
- AI5 and AI6 chip strategy
- The role of chips in Tesla’s master plan
- Manufacturing partners: TSMC and Samsung
- Conclusion: A unified heart for Tesla’s ecosystem
Introduction: A new focus after halting Dojo
Tesla has revealed that it is conducting an internal evaluation to integrate thousands of AI5 chips to train its next-generation artificial intelligence models. This evaluation appears to have yielded positive results, coming after the company’s decision in August to halt the controversial Dojo chip design project and focus all its resources on a single, unified chip architecture.
Musk’s announcement: An “epic” chip on the way
On September 7, Tesla CEO Elon Musk posted on social media that he had just completed an exceptional design review with the Tesla AI5 chip design team, stating that this chip will become an “epic” product. He added that the subsequent AI6 chip is expected to be the most prominent AI chip yet. Musk believes that AI5 could be the best inference chip for models with fewer than 250 billion parameters, because it “has the lowest silicon cost and the highest performance-to-power ratio.”
Why was the Dojo project halted?
Previously, Tesla adopted a dual-track parallel chip development strategy of “training + inference,” where Dojo was used for training and the FSD chip was used for in-vehicle inference. Musk revealed that Tesla decided to shut down the Dojo project because it did not make sense to distribute resources across two different AI chip designs. His latest statement confirms that the shift to a single architecture means all of Tesla’s chip talent will be concentrated on creating this exceptional chip.
AI5 and AI6 chip strategy
The AI5 and AI6 chips developed by Tesla are expected to strongly support its artificial intelligence training and autonomous driving. Reports indicate that AI5 is a transitional or primary chip for specific scenarios, while AI6 will serve as the “unified heart” of Tesla’s future AI system. AI5 will focus on training compute clusters for vehicle inference (such as autonomous driving) and is expected to enter mass production by the end of 2026.
The role of chips in Tesla’s master plan
Tesla’s self-developed chips are a key step in its “Master Plan,” which is expected to bring significant performance improvements and cost optimizations to Tesla’s future core products. It will also provide a greater advantage in hardware and software integration, reduce dependence on external suppliers, and provide a computing power foundation for the rapid iteration of autonomous driving technology and robotics products.
Manufacturing partners: TSMC and Samsung
The AI5 chip is scheduled to be produced by Taiwan Semiconductor (TSMC), while the AI6 will be manufactured by Samsung Electronics. The AI6 chip will first be applied in Tesla’s robotaxi, the Cybercab, and the humanoid robot, Optimus, and may later expand to AI data centers, challenging the leadership of companies like Nvidia in this space.
Conclusion: A unified heart for Tesla’s ecosystem
The progress on the AI5 chip and the unification of chip design efforts represent a crucial strategic step for Tesla. By concentrating its talent and resources on a single architecture, the company aims not only to create a best-in-class chip, but also to build an integrated and robust foundation for its future ecosystem spanning vehicles, robots, and data centers. Musk’s description of the chip as “epic” indicates his high confidence that this focused approach will pay off.
