With artificial intelligence (AI) and HPC solutions becoming more ubiquitous, organizations are looking for high-powered computing solutions that also address thermal management concerns.

The NVIDIA GB200 NVL72 is an innovative solution that meets the demands of both AI and HPC workloads. Designed for large-scale AI and HPC workloads, this liquid-cooled server rack has 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace™ CPUs to offer a large amount of computing power. What’s more, the NVIDIA GB200 NVL72 tray has direct-to-chip liquid cooling for even the most intensive thermal management requirements within an AI/HPC environment. The QoolRack Standalone solution adds significant thermal management capabilities.

Additionally, the NVIDIA GB200 NVL72 works with the MGX product line. This is a flexible design framework based on NVIDIA’s Modular GPU Accelerated architecture that enables flexible and scalable solutions for various AI and HPC workloads. MGX enables different NVIDIA components like CPUs and GPUs to work together in a way that can be easily updated or changed as new technology comes out. The modular architecture also allows users to mix and match different parts depending on what they need, making it easier for organizations to build for specific AI and HPC tasks.