A brief 20-second video on Twitter (X) provided a first look at this groundbreaking AI training supercluster, located within Tesla's expansive 10-million-square-foot Giga Texas facility in Austin. The Cortex cluster is composed of 100,000 Nvidia H100/H200 GPUs, some of the most powerful GPUs available, crucial for training large language models (LLMs) such as OpenAI's GPT-4 and Google's PaLM 2.
Cortex is more than just a supercomputer - it’s a giant leap into the future. The cluster is so vast that it needed an entire extension to be built onto the Giga Texas facility. And that’s not all; a colossal cooling system is also under construction. Imagine running 40,000 window A/C units simultaneously - that’s the kind of power Cortex will need, starting with 130 MW of cooling and ramping up to over 500 MW in just 18 months.
This ambitious project is part of Musk's broader efforts in supercomputing, running parallel to a multibillion-dollar AI supercomputer cluster for xAI, his artificial intelligence company. With an initial order of 100,000 Nvidia H100 GPUs, this supercomputer is set to become one of the world’s largest AI clusters, powering the next-gen GrokAI chatbot for X’s premium users. Dell and Supermicro are teaming up to provide servers for this powerhouse, as Dell’s CEO Michael Dell recently confirmed on Twitter.
The Cortex AI supercomputer is crucial for advancing two of Tesla's most strategic products: Full Self-Driving (FSD) technology and the Optimus humanoid robot. The success of Tesla's proposed Robotaxi network, a key element of Musk's Master Plan, Part Deux, depends heavily on the performance of FSD. Additionally, the Optimus robot holds the potential to transform multiple industries, including manufacturing. Cortex, with its massive computational power, is poised to play a pivotal role in these developments, driving Tesla's innovation and leadership in AI and robotics.
