GPU as a Service

Hikube provides access to NVIDIA accelerators via GPU Passthrough, enabling hardware-accelerated workloads (AI/ML, rendering, HPC).

Usage Modes

Mode	Description	Use Case
GPU on Kubernetes	Dedicated GPU node group in a managed cluster	ML training, inference, batch processing
GPU on VM	GPU attached directly to a virtual machine	Native CUDA, graphics rendering, legacy HPC

Provisioning Guides

GPU on Kubernetes

Provision a GPU node group in a managed Kubernetes cluster with the GPU Operator.

KubernetesGPU Operator

GPU on VM

Attach an NVIDIA GPU to a KubeVirt virtual machine via GPU Passthrough.

VMPassthrough

Overview

Architecture, available GPU types, and billing model.