Elon Musk has given the public a glimpse into Tesla’s ambitious Cortex supercomputer cluster at the Giga Texas facility. In a brief video shared on the social media platform X, Musk showcased an area of Cortex still under construction, revealing an impressive array of Nvidia H100 and H200 GPUs.
According to Musk, Cortex is set to house around 100,000 of these powerful GPUs, which will significantly enhance Tesla's artificial intelligence (AI) capabilities. "Here’s a video of the inside of Cortex today, the giant new AI training supercluster being built at Tesla HQ in Austin to solve real-world AI," Musk stated in his post.
The development of Tesla's AI infrastructure is crucial for advancing two of its key projects—Full Self-Driving (FSD) technology and Optimus, Tesla's humanoid robot. The success of the future Robotaxi network, first envisioned in Musk’s "Master Plan, Part Deux" back in 2016, heavily relies on the effectiveness of FSD. Meanwhile, Optimus holds the promise to transform various industries, especially in manufacturing.
Cortex is so vast that it necessitated a significant expansion of the Giga Texas facility. Musk also pointed out that the supercluster will initially require approximately 130 MW of cooling, with plans to increase that demand to over 500 MW in the next 18 months. Drone footage has revealed ongoing construction of a large cooling system tailored for Cortex.
In addition to the Giga Texas supercluster, Tesla is ramping up its AI initiatives elsewhere, such as its recent agreement with Buffalo, New York, to extend operations at Giga New York until 2034. Tesla plans to invest $500 million in a new Dojo supercomputer project in Buffalo, which will contribute further to its autonomous driving technologies.
Tesla 2024 news from teslarati
