Tesla and xAI are gearing up to launch new supercomputers featuring a staggering 350,000 Nvidia GPUs, expected to be operational within months. The expansion of Elon Musk’s Texas Tesla Gigafactory includes an AI supercomputer cluster, with Supermicro’s innovative liquid cooling technology being a key component. Charles Liang, CEO of Supermicro, took to X (formerly Twitter) to express his excitement about Musk’s use of their cooling solution for both Tesla’s and xAI’s supercomputers.
Liang and Musk, pictured together among server racks, are championing the use of liquid cooling technology in large AI data centers. Liang highlighted that Musk’s leadership in this area could potentially “preserve 20 billion trees for our planet,” referencing the environmental benefits of widespread liquid cooling adoption.
AI data centers are notorious for their high power consumption. Supermicro aims to mitigate this by promoting liquid cooling, which the company claims can reduce electricity costs by up to 89% compared to traditional air cooling. Liang’s previous Tweet emphasized Supermicro’s goal to increase direct liquid cooling (DLC) adoption from less than 1% to over 30% within a year.
Musk’s Tesla Gigafactory expansion will house 50,000 Nvidia GPUs, alongside Tesla AI hardware designed to train the Full Self-Driving feature. The new facility is notable for its massive fans used to chill the liquid cooling system, which Musk showcased in a recent X post. The supercomputer is expected to draw 130 megawatts of power initially, with plans to scale up to 500 megawatts as more AI hardware is installed. The construction is nearly complete, with deployment planned in the next few months.
Simultaneously, Musk is developing another supercomputer for xAI, which will initially use 100,000 Nvidia H100 GPUs to train GrokAI, a chatbot for X Premium subscribers. The xAI supercomputer, also liquid-cooled by Supermicro, has a planned upgrade to 300,000 Nvidia B200 GPUs next summer. Prioritizing the xAI supercomputer, Musk redirected thousands of GPUs initially intended for Tesla, causing delays in Tesla’s supercomputer project.
Both supercomputers underscore Musk’s commitment to advancing AI capabilities and sustainable cooling technologies, setting new benchmarks in the industry.