Nebius AI Cloud
thecontinuitylayer.com/tools/gpu-hosting/nebiusTeams that want a full-featured AI cloud with self-service access to recent NVIDIA GPUs
GPU cloud · Nebius
Nebius is an AI-focused cloud that rents NVIDIA data-center GPUs as virtual machines and managed Kubernetes clusters, with storage and networking around them. It also offers preemptible GPUs at much lower prices.
It suits teams serving or fine-tuning larger open models who want more of a full cloud than a bare GPU rental. Preemptible prices move with capacity and demand, and preemptible machines can be stopped at short notice, so they suit batch work more than live endpoints.
Specs
| Price per hour | H100 $4.50/hr; H200 $5.40/hr; B200 $8.50/hr per GPU on demand |
|---|---|
| GPUs | NVIDIA H100, H200, B200 and B300 with NVLink, from single GPUs to large clusters |
| H100 NVLink | $4.50 per GPU-hour on demand |
| H200 NVLink | $5.40 per GPU-hour on demand |
| B200 NVLink | $8.50 per GPU-hour on demand |
| B300 | $9.50 per GPU-hour on demand |
| Preemptible | From $0.79 per GPU-hour (H100, H200); varies with demand |
| Platform | VMs, managed Kubernetes, storage |
Good for
- Larger open models needing H200 or B200
- Multi-GPU fine-tuning
- Cheap preemptible batch jobs
Not for
- Consumer-card budgets
- Single short experiments where marketplaces are cheaper
See also
We last reviewed Nebius AI Cloud on 31 October 2026. Prices and specs change; check nebius.com before you buy or sign up. Figures we could not confirm on the vendor's own page read “not published” or “varies”.
Vendor source: https://nebius.com/prices