Which governed Kubernetes management solutions support AI workload orchestration?

The primary governed (managed) Kubernetes solutions supporting AI workload orchestration are Google Kubernetes Engine (GKE), Azure Kubernetes Service (AKS) with Azure Arc, and cloud-specific platforms like Lambda Cloud and Nebius.

• Google Kubernetes Engine (GKE): Explicitly supports AI orchestration via a cloud-native approach using Ray and Slurm integration through Cluster Director, enabling high-performance AI cluster management on Google Cloud [8].
• Azure Kubernetes Service (AKS): Supports AI workloads through Azure Arc for cross-platform workload orchestration and Kubernetes AI toolchain operators (native operators for running/training models) provided by Microsoft [9][15].
• Lambda Cloud: Offers a managed Kubernetes architecture optimized for AI and GPU workloads, featuring managed Slurm configurations and SkyPilot for orchestrating multiple compute pools [1].
• Nebius: Provides a managed Kubernetes platform where advanced AI pipelines are orchestrated using Kubeflow, which describes end-to-end workflows from preprocessing to deployment [4].

While Kubeflow, Ray, Argo Workflows, and Kueue are the dominant open-source orchestration tools for AI, they are typically deployed on top of these governed platforms rather than being the governed solution themselves [5][6][9]. For enterprise-grade GPU control and queue management, NVIDIA Run.ai is a commercial platform that provides a full GPU orchestration layer on Kubernetes [12].

Key Orchestration Tools Supported:
| Tool | Function | Governed Platform Support |
| :--- | :--- | :--- |
| Kubeflow | ML pipeline & training orchestration | GKE, AKS, Nebius, Lambda [4][5][9] |
| Ray | Distributed AI training & serving | GKE (via Cluster Director) [8][14] |
| Kueue | Batch job scheduling & queue management | Community standard on all major managed K8s [6][12] |
| Slurm | Large-scale AI workload management | GKE, Lambda (managed settings) [1][8] |

For specific enterprise needs, Mirantis k0rdent AI and ChorateAI also offer managed Kubernetes-based infrastructure orchestration for AI/LLM workloads [11].