Asus Introduces Liquid Cooling for AI Data Centers

Why It Matters

By dramatically improving cooling efficiency, ASUS’s liquid‑cooling reduces energy consumption and PUE, lowering total cost of ownership for AI supercomputers while enabling higher rack density and performance—critical as AI workloads surge. This positions ASUS as a key infrastructure provider in the rapidly expanding AI data‑center market.

Key Takeaways

• ASUS launches optimized liquid-cooling for AI/HPC data centers
• Solutions cover D2C, in-row CDU, hybrid configurations
• Partners include Schneider, Vertiv, Cooler Master, Auras Technology
• NCHC deployment achieves PUE 1.18 with NVIDIA HGX H200
• ASUS claims 2,156 #1 SPEC CPU, 248 #1 MLPerf records

Pulse Analysis

The surge in AI and high‑performance computing workloads is pushing server racks beyond the thermal limits of traditional air‑cooling. Data‑center operators face escalating power densities that translate into higher energy bills and lower reliability. ASUS’s Optimized Liquid‑Cooling Solutions respond to this pressure by moving heat directly away from CPUs, GPUs, and accelerators, enabling tighter packing of components without compromising stability. This shift mirrors a broader industry trend where liquid‑cooling is becoming a prerequisite for exascale ambitions.

ASUS differentiates its offering through a modular portfolio that includes direct‑to‑chip (D2C) cooling plates, in‑row cooling distribution units (CDUs), and hybrid configurations that blend air and liquid pathways. By partnering with infrastructure leaders like Schneider Electric and Vertiv, ASUS leverages proven power‑management and rack‑integration expertise, accelerating deployment cycles. The technical benefits are tangible: lower Power Usage Effectiveness (PUE) scores, reduced cooling‑related electricity consumption, and a slimmer total cost of ownership (TCO) for operators seeking sustainable growth. These efficiencies also align with corporate ESG goals, making liquid‑cooling an attractive proposition for environmentally conscious enterprises.

The real‑world impact is illustrated by the National Center for High‑Performance Computing’s first fully liquid‑cooled AI supercomputer in Taiwan. Featuring NVIDIA HGX H200 and GB200 GPUs, the system achieves a PUE of 1.18, setting a new benchmark for energy efficiency in AI clusters. This deployment validates ASUS’s claim of delivering high‑density, low‑PUE solutions at scale and signals to the market that liquid‑cooling is no longer a niche technology but a mainstream requirement for next‑generation AI infrastructure. As competition intensifies, vendors that can combine performance, efficiency, and partner ecosystems will likely capture the most lucrative data‑center contracts.