DC GPU Compute

GPU

GPU-Accelerated VMs

DC GPU Compute adds graphics and compute acceleration to standard virtual machines. Choose passthrough mode for near-native performance on a dedicated physical GPU — ideal for model training and rendering — or vGPU profiles that share one physical GPU across several VMs for virtual desktops and inference workloads.

Get Started View Pricing

Key Features

Passthrough mode: 1–4 full physical GPUs dedicated to a single VM

vGPU profiles: partition one GPU across several VMs for desktops and inference

Install standard vendor drivers in the guest exactly as on physical hardware

GPU VMs keep full platform protection — snapshots, backup, and monitoring

Per-hour GPU metering, billed only while attached

Architecture Overview

Source / Trigger

Core Component

Connected Service (clickable)

Pricing

Priced to your configuration

DC GPU Compute is scoped and priced against your workload requirements. Standardized public packages are being finalized.

Request Pricing

Common Use Cases

Machine learning training and inference

3D rendering and simulation

GPU-accelerated virtual desktops

Video processing pipelines

Getting Started

REST API — Create a GPU VM

curl -X POST "https://api.daakyicloud.com/v1/vms" \
  -H "Authorization: Bearer $TOKEN" \
  -d '{
    "name": "training-01",
    "vcpus": 16,
    "memory_gb": 64,
    "image": "ubuntu-24.04",
    "gpu": {"mode": "passthrough", "count": 2},
    "subnet_id": "subnet-a1b2c3"
  }'

# { "id": "vm-9e2c", "gpu": "2x passthrough",
#   "status": "provisioning" }