Elon Musk's vision for Tesla goes beyond just being an automaker, with the company aiming to become a leading AI company that can make cars drive themselves. Central to this mission is Dojo, Tesla's custom-built supercomputer designed to train its Full Self-Driving (FSD) neural networks. While FSD isn't yet fully self-driving, requiring an attentive human behind the wheel, Tesla believes that with more data, compute power, and training, it can cross the threshold to full autonomy.

Dojo has been a key focus for Musk, with the executive teasing the supercomputer throughout 2024. The company has made significant progress, with the installation of the first Dojo cabinet and the testing of 2.2 megawatts of load testing. Tesla has also set a target date of completing a full Exapod cluster by Q1 2023 and plans to build a total of seven Exapods in Palo Alto.

In 2021, Tesla officially announced Dojo at its first AI Day, introducing its D1 chip, which will power the Dojo supercomputer alongside Nvidia's GPU. The company also released a Dojo Technology whitepaper, outlining a technical standard for a new type of binary floating-point arithmetic used in deep learning neural networks. By 2022, Tesla had installed the first Dojo cabinet and was testing its capabilities, with Musk noting that Dojo "has the potential for an order of magnitude improvement in the cost of training" and could become a sellable service to other companies.

In 2023, Musk told investors that Dojo was already online and running tasks at Tesla data centers, with the company projecting that its compute power would be among the top five in the world by February 2024. However, it's unclear if this target was met. Tesla also started production of Dojo, with Musk saying the company would spend over $1 billion on the project through 2024.

In 2024, Tesla announced plans to spend $500 million to build a Dojo supercomputer in Buffalo, with Musk downplaying the investment, noting that it was equivalent to a 10k H100 system from Nvidia. The company also revealed that its next-generation training tile, the D2, was already in production, and Musk noted that the rear portion of the Giga Texas factory extension would include the construction of "a super dense, water-cooled supercomputer cluster."

However, Tesla has faced challenges in scaling up Dojo, with Musk acknowledging that the project is high-risk, high-reward. The company has also faced supply chain issues with Nvidia hardware, with Musk saying that demand is "so high that it's often difficult to get the GPUs." This has led Tesla to focus more on Dojo, with Musk saying that the company sees a path to being competitive with Nvidia with the supercomputer.

In a recent update, Musk revealed that Tesla's current vehicles may not have the right hardware for the company's next-gen AI model, requiring an upgrade to the vehicle inference computer. The company has also shifted focus to its new AI training supercluster, Cortex, which is made up of roughly 50,000 H100 Nvidia GPUs. During Tesla's Q4 and full-year 2024 earnings call, the company noted that Cortex had enabled the rollout of FSD V13, boasting major improvements in safety and comfort.

While there have been no updates on Dojo in 2025, Tesla's continued investment in AI and its focus on Cortex suggest that the company remains committed to achieving full self-driving autonomy. As the company continues to push the boundaries of AI technology, it will be interesting to see how Dojo and Cortex contribute to this mission.