Tesla invests $500 million in the construction of the Dojo supercomputer in New York
Tesla is preparing to build the next-generation Dojo supercomputer at its Gigafactory in Buffalo as part of a $500 million investment.
The supercomputer is designed to process the huge amount of data coming from Tesla cars and train artificial intelligence (AI) systems for autonomous driving and other applications. It is expected to become one of the most powerful computing clusters in the world and surpass the current leader, NVIDIA.
New York Governor Kathy Hochul announced this during a hearing on the state’s economic development, saying that the Tesla project will create hundreds of jobs and contribute to the development of the region’s innovation ecosystem. She also said that the state would support a separate project to build an AI supercomputer at the State University of New York (SUNY) with $275 million and more than $400 million in public and private investment.
Tesla CEO Elon Musk confirmed these plans on the X platform (blocked in Russia). However, he also clarified that the $500 million figure refers only to a smaller system based on NVIDIA chips, and the cost of the Dojo supercomputer will be much higher. He said that the stakes for competitiveness in the AI field are at least several billion dollars a year.
The Tesla Gigafactory in New York City currently produces solar panels, solar roofs, and Supercharger components, and is home to the Autopilot data analytics team. Hochul did not name the timeframe for installing the Dojo supercomputer. The governor also said that it will be a turning point for the state’s AI industry, which is expected to grow from $150 billion to $1.3 trillion in the next few years. This statement was made after Tesla’s Dojo project manager Ganesh Venkataramanan left the company in December 2023. In June 2023, Musk announced that Dojo was already online and performing useful tasks for several months.
As previously reported, Morgan Stanley analysts said that Tesla has developed an advanced supercomputer architecture that pushes new boundaries in custom silicon and could provide Tesla with an asymmetric advantage in the $10 trillion TAM.
Tesla’s Dojo supercomputer will use AI ASIC chips, which are reportedly more efficient and powerful than NVIDIA’s A100 GPUs. Tesla has used NVIDIA chips for its existing supercomputer, which is believed to have more than 10 thousand. units. However, NVIDIA chips are experiencing a global shortage and growing demand, which may have prompted Tesla to develop its solution.
