Elon Musk's huge liquid-cooled Gigafactory AI supercomputers get praise from Supermicro CEO

Elon Musk is building a gigantic AI supercomputer at Tesla's Texas Gigafactory, with 350,000 liquid-cooled AI GPUs that Supermicro CEO Charles Liang recently praised. Check out his post on X below:

The Supermicro CEO was photographed next to Elon Musk with some AI server racks, where he said the duo will "lead the liquid cooling technology to large AI data centers". Liang estimates that Musk's moving into the world of liquid-cooled AI supercomputers "may lead to preserving 20 billion trees for our planet" if more AI data centers moved to liquid cooling.

Data centers consume monumental amounts of power, with Supermicro hoping to reduce that through the use of liquid cooling, with the company claiming that direct liquid cooling could reduce electricity costs of cooling infrastructure compared to air cooling by 89%. In a previous tweet, the Supermicro CEO said the company's goal is "to boost DLC (direct liquid cooling) adoption from <1% to 30%+ in a year".

Musk's existing Gigafactory in Texas will be powered by over 50,000 NVIDIA AI GPUs and Tesla AI hardware in order to train Tesla's leading Full Self Driving (FSD) feature for its electric vehicle family.

• Read more: Elon Musk to buy over $10 billion worth of NVIDIA's new B200 AI GPUs for xAI supercomputer

Tesla's new Gigafactory supercomputer cluster isn't the same as Musk's supercomputer cluster for X and xAI, which is under construction right now. The new xAI supercomputer is more well known, with Musk ordering a hefty 100,000+ NVIDIA H100 AI GPUs, which will train GrokAI, the AI chatbot for X that's available to Premium subscribers.

• Read more: NVIDIA CEO proud to announce Dell's new AI factory is powered by Blackwell AI GPUs
• Read more: NVIDIA, Foxconn expect results this year for AI factories, smart manufacturing, AI smart EVs
• Read more: NVIDIA's next-gen B200 AI GPU uses 1000W of power per GPU, drops in 2025, confirmed by Dell
• Read more: Dell teases an Arm-based processor from NVIDIA inside of a next-gen AI PC in 2025

NVIDIA CEO Jensen Huang talked about Dell being a long-term partner in May 2024, he said at the time: "I'm here, I'm here to announce this Dell AI factory with NVIDIA going to market starting today. In order to do this, just think about what an AI factory is. It has CPUs and GPUs".

Jensen continued: "It has networking switches. Incredible amounts of software in systems connected to storage and the miles of, the miles of cables alone necessary to build up one of these things. Well we. . . you know, Michael, turned it into an easy bit. And so that, just as Dell did in early days with PCs, they're now doing this with AI factories".