NVIDIA-Certified Storage

This Enterprise Reference Architecture is designed to support high performance storage solutions that can keep pace with GPU intensive AI workloads. As a guideline, the architecture allocates approximately 12.5 Gb/s per GPU, scaling linearly as the cluster grows—for example, a 16 GPU cluster would require around 200 Gb/s of aggregate storage bandwidth. For validated storage systems and deployment guidance tailored to your cluster size, refer to the NVIDIA-Certified Storage program, which integrates seamlessly with this reference architecture.

The NVIDIA-Certified Storage program provides a comprehensive validation framework that ensures storage platforms meet stringent performance, quality, and interoperability standards for AI workloads. Partner solutions undergo extensive functional and performance testing to guarantee predictable deployment, energy efficiency, and operational consistency for end-to-end AI factories.