Tesla (NASDAQ: TSLA) launched its official Twitter account, @Tesla_AI, in May with a focus on sharing updates about the company’s artificial intelligence (AI) progress related to its FSD software and upcoming humanoid robot, Teslabot.
On June 21, the dedicated AI account made another significant announcement. According to the published graphic, Tesla’s highly anticipated Dojo supercomputer, regarded as the largest electric vehicle (EV) manufacturer globally, is scheduled to enter production in July 2023.
And will be trained on enormous amounts of compute pic.twitter.com/BsmG9Vse6I
— Tesla AI (@Tesla_AI) June 21, 2023
By referring to the chart, it can be observed that Tesla plans to continuously enhance its computing power until it ranks among the top 5 most powerful supercomputers worldwide by early 2024. The company aims to achieve this by utilizing its extensive collection of real-world driving videos, which it has been accumulating for several years. Tesla expects its compute capacity to reach 100 exaflops by October 2024.
What is Dojo?
Dojo is Tesla’s proprietary custom-built supercomputer platform specifically developed for AI machine learning. It will be utilized for training neural networks using video data obtained from Tesla’s own vehicles. This supercomputer is expected to play a crucial role in advancing Tesla’s computer vision technology, which serves as the foundation for its self-driving solution.
Tesla is building the foundation models for autonomous robots pic.twitter.com/VUES9jU3ze
— Tesla AI (@Tesla_AI) June 21, 2023
While Elon Musk’s EV manufacturing company already possesses a powerful NVIDIA GPU-based supercomputer, the newly introduced Dojo employs chips and a complete infrastructure designed by Tesla itself.
In a comprehensive Twitter thread, Tesla’s AI account described the automaker’s efforts in building fundamental models for autonomous robots. Additionally, the thread discussed the integration of the company’s multi-model neural networks into its fleet of electric cars.
