NVIDIA Cloud Native Technologies

Type: Platform Tags: NVIDIA, cloud native, Kubernetes, GPU Operator, Container Toolkit, device plugin, MLOps Related: NVIDIA-GPU-Operator, NVIDIA-Network-Operator, Nsight-Cloud, NVIDIA-AI-Cluster-Runtime, KAI-Scheduler, NVIDIA-NVSentinel, NVIDIA-Container-Toolkit, NVIDIA-DCGM, NVIDIA-DGX-SuperPOD, NVIDIA-Cloud-Accelerator-NCX, NVIDIA-AI-Enterprise Sources: https://docs.nvidia.com/datacenter/cloud-native/index.html, https://developer.nvidia.com/nsight-cloud Last Updated: 2026-04-29

Summary

NVIDIA Cloud Native Technologies is the documentation hub for running GPU-accelerated containers and Kubernetes clusters on NVIDIA GPUs. It includes the operational building blocks that expose GPUs to containers and automate the NVIDIA GPU software stack.

Detail

The docs hub points to cloud-native components such as GPU exposure on Kubernetes nodes and GPU-enabled container runtime support. In practice, the core pages in this area include NVIDIA-GPU-Operator, NVIDIA-Network-Operator, NVIDIA-Container-Toolkit, device plugins, DCGM integration, and Kubernetes deployment patterns.

This is a hub page rather than a replacement for the component pages. It should make it easier to answer questions about “NVIDIA Kubernetes stack” or “cloud-native NVIDIA GPU infrastructure” from the wiki. Newer NCX cloud-native components such as NVIDIA-AI-Cluster-Runtime, KAI-Scheduler, and NVIDIA-NVSentinel extend this layer into validated cluster recipes, scheduling, and remediation. Nsight-Cloud adds the cloud-native profiling path for Kubernetes and containerized workloads.

Connections

• NVIDIA-GPU-Operator - automates driver, device plugin, container toolkit, DCGM, and MIG components.
• NVIDIA-Network-Operator - automates NVIDIA networking drivers, RDMA, SR-IOV, secondary networks, and DOCA-OFED in Kubernetes.
• Nsight-Cloud - cloud-native Nsight profiling toolkit for Kubernetes, containers, Operator, and Streamer workflows.
• NVIDIA-AI-Cluster-Runtime - publishes validated recipes for NVIDIA-accelerated Kubernetes runtimes.
• KAI-Scheduler - GPU-aware Kubernetes scheduler for AI workloads.
• NVIDIA-NVSentinel - Kubernetes-native GPU fault detection and remediation.
• NVIDIA-Container-Toolkit - enables container runtimes to use NVIDIA GPUs.
• NVIDIA-DCGM - GPU metrics and diagnostics for cloud-native monitoring.
• NVIDIA-DGX-SuperPOD - target cluster infrastructure for Kubernetes GPU workloads.
• NVIDIA-Cloud-Accelerator-NCX - cloud partner architecture includes container-service components.

Source Excerpts

• NVIDIA Cloud Native Technologies docs describe GPU exposure on cluster nodes and running GPU-enabled containers in Kubernetes clusters.