ARS-511GD-NB-LCC-01-G2

Data Center-Level Performance in a Local AI Supercomputer

Built on the same GB300 architecture that powers large AI factories around the world, the Super AI Station delivers up to 20 PFLOPS of FP4 AI performance and up to 3x the fine tuning and inference performance of H200 to support autonomous agents, AI development, AI inference, and data science. A 748GB combined coherent memory pool supports trillion-parameter models locally, allowing always-on agents to run 24/7, generating up to 1200tps at no additional token cost. Full local AI development capabilities eliminates reliance on cloud-based AI services, reducing security risks for proprietary data and offloading cloud subscription costs. The system offers native support for the full suite of NVIDIA AI Enterprise (NVAIE) software stack, including software tools, libraries, and frameworks for enterprise AI deployments.