AI Reference Design for NVIDIA 1 SU DGX H100 SuperPODs

Application Brief

As generative AI reshapes the demands of the modern data center, infrastructure must evolve to support massive compute density, rising power loads, and precision cooling. In this AI Reference Design, we detail the Legrand solution offering that supports the implementation of a NVIDIA DGX H100 SuperPOD.

Whether you're a hyperscaler, integrator, or enterprise, this comprehensive reference architecture provides all the building blocks you need—from racks and power to connectivity and cooling—to deploy scalable AI clusters with confidence.

What You'll Find Inside:

• A line-by-line matchup of NVIDIA requirements to Legrand solutions
• Bill of Materials to streamline planning and procurement
• Layout recommendations for linear and across-the-aisle configurations
• Intelligent thermal monitoring and power distribution strategies for AI racks