AI and HPC growth drives demand for liquid cooling and high-performance GPU servers

The rapid expansion of artificial intelligence (AI) and high-performance computing (HPC) is significantly increasing the power demands and heat generation of data centers. As modern AI clusters often exceed 50 kW per rack—with some advanced solutions surpassing 100 kW—traditional air cooling is reaching its physical and economic limits.

Liquid cooling technologies, such as Direct-to-Chip (D2C) cooling, are becoming a necessity. These systems use liquid to transport thermal energy more efficiently than air, utilizing cold plates placed directly on high-heat components like CPUs and GPUs to maintain stability and energy efficiency.

To support these workloads, new hardware solutions like the INNO3D AGS-6220V2 6U rack-mount server are being developed. This system features a dual-processor platform based on 5th generation Intel Xeon Scalable processors and is designed to support high-end accelerators, such as NVIDIA RTX PRO 6000 Blackwell Server Edition cards, to handle intensive AI model training and data center operations.