VISI

What is Run:ai

NVIDIA Run:ai is a Kubernetes-based AI workload management platform for scheduling AI jobs and optimizing GPU resource allocation.

Run:ai on DGX Cloud is the managed-service version: NVIDIA provisions and maintains all infrastructure. Customers interact via the Run:ai UI and CLI.

Each compute node in DGX Cloud = 8× NVIDIA H100 GPUs. Customer gets a Realm (tenant) in the shared control plane.

Back to moduleNext: Run:ai organization and roles

Study tools

Mark this unit complete when you finish it.