NVIDIA Run:ai TRIAL

Run:ai now combines memory-enforced GPU fractioning with Workload Map reuse across Kubernetes AI workload types. GPU failure domains and custom-CRD status propagation still require workload-specific validation.

Next stepWithin one day, test GPU sharing and custom-CRD mapping separately; record resource accounting, status propagation and failure spillover.

Judgment reviewed 2026-11-03 · Basis: NVIDIA extends Run:ai with GPU controls and Workload Map reuse