Penguin Solutions announced its ComputeAI NML72-VR200. This pre-validated, liquid-cooled rack-scale platform targets artificial intelligence workloads. It utilizes NVIDIA's Vera Rubin NVL72 architecture. The system simplifies deployment and reduces integration complexity for enterprises and AI cloud providers.

The platform integrates 72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs. It also includes NVIDIA's NVLink, networking, and BlueField DPU technologies. The system is optimized for demanding AI tasks, including large-scale model training, high-concurrency agentic applications, and long-context reasoning. It is part of Penguin's Full-Stack AI Factory Platform.

By offering a fully integrated and pre-validated system, Penguin aims to help organizations accelerate their time to production for next-generation AI initiatives. The ComputeAI NML72-VR200 is available now.