NVIDIA open-sources Run:ai under Apache 2.0 for Kubernetes GPU orchestration

NVIDIA has released the full Run:ai platform as open source under Apache 2.0 on GitHub. The release splits into KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map; the community edition excludes enterprise support.

• Monorepo ships control plane, cluster services, CLI and an MCP server
• Local dev builds a 5-node Kind cluster with 8 simulated Tesla-A10 GPUs
• GPU Fractioning uses CUDA 13.4-native memory isolation per workload
• Model Streamer loads models 3–6x faster on 8 H100 VMs