ASUS introduces optimized liquid-cooling solutions

ASUS announced Optimized Liquid‑Cooling Solutions and a strategic partner framework to address rising thermal, power and density challenges in next‑generation Artificial Intelligence (AI) and High performance computing (HPC) data centers.

The company framed the offering as a response to compute density and power demands that exceeded traditional air cooling, and said the solutions would provide thermal management for data centers built around NVIDIA Vera Rubin NVL72 systems.

The portfolio covered direct‑to‑chip (D2C), in‑row CDU‑based cooling, and hybrid configurations and was described as supporting heat dissipation from high‑performance CPUs, GPUs, and accelerator‑dense racks to reduce energy consumption, lower Power Usage Effectiveness (PUE), and optimize Total Cost of Ownership (TCO). ASUS cited 2,156 No. 1 SPEC Central Processing Unit (CPU) records and 248 No. 1 MLPerf results.

ASUS outlined a partner framework that included Schneider and Vertiv, with precision components from Auras Technology and Cooler Master. As a deployment example, ASUS delivered a dual‑compute system for the National Center for HPC, National Institutes of Applied Research (NCHC, NIAR) in Taiwan that combined a Nano4 NVIDIA HGX H200 cluster and an NVIDIA GB200 NVL72 system, and said the deployment was Taiwan’s first fully liquid‑cooled AI supercomputer of this architecture; ASUS said the system used Direct Liquid Cooling (DLC) and achieved a PUE of 1.18 and that it was designed and engineered by ASUS from the ground up.

ASUS said it would appear as a Diamond Sponsor (Booth #421) at NVIDIA GTC 2026, March 16–19 in San Jose, and said its servers were available worldwide.