NVIDIA Mission Control 3.0 introduces a modular, API-driven architecture for AI factory management with three major enhancements: multi-org isolation using virtualized infrastructure with VXLAN/PKey network segmentation, intelligent power orchestration that makes power a first-class scheduling primitive (achieving 85% power usage with only 7% throughput loss), and AI-powered predictive anomaly detection via NACPS using graph-based cluster models. The release shifts the operational KPI from GPU utilization to token production per watt and per rack, enabling automated remediation across Slurm and Kubernetes workloads.
Table of contents
Flexible software that unlocks velocityIsolation in a multi-tenant worldPower: The invisible constraintFrom dashboards to real-time decisionsA different kind of KPI: Utilization vs. token productionGet started with Mission ControlSort: