January 19, 2026- Tesla is restarting work on its custom AI chip initiative, Dojo3, after CEO Elon Musk confirmed progress on its advanced AI5 chip design. The move signals a renewed push into high-performance computing as Tesla sharpens its focus on full self-driving (FSD) capabilities and AI infrastructure.
In a post on X, Musk said development of the AI5 chip had reached a milestone that justified relaunching the broader Dojo3 project. He also described the chip as part of “the highest volume chips in the world,” suggesting large-scale deployment plans across Tesla’s fleet and data centres.
Dojo, Tesla’s in-house AI supercomputer project, aims to train neural networks at scale. It’s critical to Tesla’s long-term strategy to lead in autonomous driving and AI-powered robotics. Musk previously paused parts of the project in 2025 to rework design issues and shift talent to other AI-related priorities. Now that a new generation of chips is on the horizon, the project is back in motion.
Musk’s announcement also included a call for engineers and chip designers to join the effort, indicating Tesla is scaling hiring for its AI hardware division. Industry observers see this as a direct challenge to Nvidia, which currently dominates the AI chip market with its GPUs.
The decision comes at a time when competition in AI hardware is heating up. Amazon, Meta, and Google are all investing in custom chips to reduce their reliance on third-party suppliers and optimise performance for specific workloads. Tesla’s strategy echoes this trend, aiming to integrate vertically across both hardware and software.
With Dojo3, Tesla is positioning itself not just as an EV maker, but as a player in the broader AI infrastructure race. Investors will be watching closely as execution on this complex project unfolds.
