ASUS Launches Advanced Liquid-Cooled AI Servers
ASUS Launches Advanced Liquid-Cooled AI Servers
As of March 2026, ASUS has officially shifted the landscape of computing by launching a series of Advanced Liquid-Cooled AI Servers.
Moving away from traditional air cooling, these new systems are designed to handle the extreme power density required for next-generation AI.
At the heart of this innovation is the ASUS AI POD, a powerhouse built on the NVIDIA Vera Rubin architecture, capable of supporting trillion-parameter LLM training with 100% liquid cooling.
By utilizing methods like Direct-to-Chip cooling, ASUS enables hardware to run at peak capacity without thermal throttling.
A standout achievement is the system's energy efficiency; these setups have reached a Power Usage Effectiveness (PUE) of 1.18.
This is a critical milestone, as it proves that these 'AI Factories' can achieve massive compute density while drastically cutting operational costs and carbon emissions.
Through partnerships with industry leaders like Vertiv, ASUS is transforming high-performance computing from an experimental endeavor into a reliable, sustainable, and purpose-built infrastructure for the future of artificial intelligence.
