The Inventec Artemis rack, powered by the NVIDIA GB200 NVL72 system, is a fully integrated, rack-level solution designed to accelerate the most demanding AI workloads. Built on the NVIDIA® MGX™ platform, it features 36 NVIDIA Grace™ CPUs paired with 72 Blackwell GPUs, all interconnected through the latest NVIDIA NVLink™ technology, which connects every node into a unified, massive-scale compute engine to deliver the performance of a single, powerful GPU at rack level. To ensure stability and sustained performance under the most intense AI training and inference demands, the entire rack is equipped with an advanced liquid cooling system that efficiently manages thermal loads. With this architecture, organizations can tackle large-scale generative AI workloads with confidence, efficiency, and the scalability needed to unlock the next frontier of AI innovation.