NVIDIA has released the Run:ai platform’s source within its DSX AI Factory ecosystem. The release covers the integrated platform and independently adoptable components: KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map.
The platform allocates GPUs and team quotas on Kubernetes across data processing, training and inference. Source is available under dsx-ai-factory; an NGC Community Product includes a Helm chart. NVIDIA says it requires no paid Run:ai subscription and includes no enterprise support.
Source-code licenses and NGC acceptance terms are separate layers. KAI Scheduler’s stated Apache 2.0 license must not be generalized to every component or distribution without checking. Commercial offerings continue.
Shared-compute control is becoming an open infrastructure layer separate from model selection.