Supermicro - NVIDIA GB200 NVL72 - Rack-Scale AI Supercomputer

The Supermicro NVIDIA GB200 NVL72 is a liquid-cooled, rack-scale AI supercomputer designed for the most demanding generative AI and LLM training workloads. It integrates 72 Blackwell GPUs and 36 Grace CPUs into a single, massive compute fabric using fifth-generation NVLink technology. - Features 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace CPUs in a single rack configuration. - Delivers up to 1.4 exaflops of AI performance and 30TB of fast memory capacity. - Fifth-generation NVLink provides 130TB/s of aggregate bandwidth for seamless GPU communication. - Advanced liquid-cooling infrastructure for maximum thermal efficiency and compute density. - Integrated NVIDIA BlueField-3 DPUs for accelerated networking, storage, and security. - Supports massive-scale LLM inference with up to 30x faster performance than previous generations. - Optimized for trillion-parameter model training and real-time generative AI inference. - Fully integrated rack-level solution including power distribution, cooling, and networking. - Reduces energy consumption by up to 25x compared to traditional air-cooled alternatives. - Designed for rapid deployment in hyperscale and enterprise data center environments.