Tesla and SpaceX CEO Elon Musk recently provided an update on the timeline for the next iterations of Tesla’s Dojo supercomputer, a critical component in the company’s pursuit of autonomous driving.
During an appearance on the All-In podcast, Musk revealed that the Dojo 2 supercomputer is expected to enter volume production by the end of 2025. This iteration is anticipated to be comparable in performance to NVIDIA’s B200 training system, a prominent player in the AI hardware market. However, Musk emphasized that the true potential of Dojo will only be realized with the third major iteration, which is expected to arrive in late 2026.
Musk’s comments highlight a common trend in technology development: it often takes several iterations for a technology to reach its peak performance. In this case, the Dojo supercomputer, designed to train the neural networks used in Tesla’s Full Self-Driving (FSD) technology, is expected to mature significantly over the next few years.
Tesla’s commitment to the Dojo project is substantial. The company has already announced an investment of over $1 billion by the end of 2024, demonstrating its dedication to advancing AI capabilities. While Tesla relies on NVIDIA for some AI processing, the development of Dojo signifies a long-term strategy to develop in-house AI expertise.
This significant investment in Dojo reinforces the importance of AI in Tesla’s future. The supercomputer’s role in training the neural networks powering FSD is crucial to achieving Tesla’s vision of fully autonomous vehicles. As Dojo evolves, it will be interesting to see how Tesla leverages its advanced AI capabilities to further innovate in the autonomous driving space.
