OVHcloud
European cloud with GPU instances and managed AI services
OVHcloud is under the market median on 1 of its 7 GPUs today.
OVHcloud is a European cloud provider that designs its own servers and operates its own data centers. Its Public Cloud offers hourly-billed NVIDIA GPU instances alongside managed AI services for notebooks, training jobs, model deployment, and token-priced inference through AI Endpoints.
At a glance · prices collected Nov 11, 2026
What's good about OVHcloud
- European provider with data centers in France, Germany, the UK, and Poland, relevant for GDPR and data-residency requirements
- Hourly and monthly billing on the same GPU flavors
- Both GPU instances and a per-token inference API from one account
- Published list prices for every GPU flavor
OVHcloud GPU prices
7 GPU models across 19 configurations. Rank compares OVHcloud's cheapest on-demand listing for each GPU with every provider's cheapest.
| GPU | OVHcloud | -60%median+100% | Rank | vs. median | Report |
|---|---|---|---|---|---|
| H200 SXM 8× | $6.20/hr | #35 of 44 providers | +36% | ||
| L40S 1× | $1.80/hr | #29 of 39 providers | +65% | ||
| L4 1× | $1.00/hr | #16 of 21 providers | +24% | ||
| Tesla V100 1× | $0.77/hr | #11 of 18 providers | +79% | ||
| H100 PCIe 1× | $2.99/hr | #8 of 14 providers | +4% | ||
| A10 1× | $1.00/hr | #3 of 13 providers | -17% | ||
| Quadro RTX 5000 1× | $0.60/hr | #2 of 3 providers | at median |
All configurations
Prices last updated November 11, 2026
| A10 | 1×2×4× | 1×2×4× |
| H100 PCIe | 1×2× | 1×2× |
| H100 PCIe | 4× | 4× |
| H200 SXM | 8× | 8× |
| L4 | 1×2×4× | 1×2×4× |
| L40S | 1×2×4× | 1×2×4× |
| Quadro RTX 5000 | 1×2×3× | 1×2×3× |
| Tesla V100 | 1× | 1× |
| Tesla V100 | 2×4× | 2×4× |
What a run costs on OVHcloud
24 hours at today's cheapest listing for each GPU, at the size it is listed.
8× H200 SXM for 24 hours
At $6.20/GPU-hr
$1,189.44
$317.80 over the market median rate
1× L40S for 24 hours
At $1.80/GPU-hr
$43.20
$17.04 over the market median rate
1× L4 for 24 hours
At $1.00/GPU-hr
$24.00
$4.68 over the market median rate
The market median is every provider's cheapest on-demand price for that GPU.
OVHcloud LLM API prices
Pay-per-token pricing, per 1M tokens.
Prices last updated November 11, 2026
| Model | Input/1M | Output/1M |
|---|---|---|
| $0.050 | $0.180 | |
| $0.070 | $0.260 | |
| $0.090 | $0.470 | |
| $0.110 | $0.110 | |
| $0.120 | $0.180 | |
| $0.470 | $3.19 | |
| $0.470 | $3.19 | |
| $0.710 | $4.25 | |
| $1.01 | $1.01 |
OVHcloud pricing by model
Input, output, and batch rates, plus alternatives, for one model at a time.
Billing & terms
- Hourly
- Pay-as-you-go billing per hour of instance use
- Monthly
- A fixed monthly rate per instance, lower than the equivalent hourly usage
- Savings Plans
- Commit to Public Cloud resources for 1 to 36 months for up to 30% off instance rates
- Per token
- AI Endpoints bills input and output tokens separately per model
- Regions
- GPU flavors run in Gravelines (France), Beauharnois (Canada), Frankfurt (Germany), London (UK), and Warsaw (Poland), with availability varying by GPU type
- Support
- Standard, Premium, Business, and Enterprise support plans, plus documentation and community forums
Watch out for
- GPU lineup centers on H200, H100, L40S, and L4; no Blackwell instances on the public price list
- Most GPU flavors are available in only one or two regions
- Part of the GPU list is older V100 and V100S hardware
- The A100 form factor (PCIe or SXM) is not stated on the price list
Features
- Cloud GPU instances
- Hourly-billed virtual machines with NVIDIA H200, H100, L40S, L4, A10, V100S, V100, and Quadro RTX 5000 GPUs in 1, 2, 4, and 8-GPU flavors
- AI Endpoints
- OpenAI-compatible inference API for open-weight language, vision, embedding, and speech-to-text models, billed per token
- AI Notebooks, Training, and Deploy
- Managed services that run Jupyter notebooks, containerized training jobs, and model-serving apps on GPUs billed by the minute
- Own hardware and data centers
- OVHcloud builds its servers and runs its own data centers, with water cooling across its fleet
- Managed Kubernetes
- Managed Kubernetes Service that can schedule workloads on GPU instance node pools
- Compliance certifications
- ISO 27001, ISO 27701, HDS (health data hosting), and SecNumCloud certifications, with SecNumCloud-qualified offerings
Services
Cloud GPU
OpenStack-based GPU virtual machines billed hourly or monthly
AI Endpoints
Serverless inference API for open-weight models, billed per input and output token
Company
Published by OVHcloud to identify the company behind this listing.
- Headquarters
- 🇫🇷 Roubaix, France
Getting started
Step 1
Create an account
Sign up for an OVHcloud account and create a Public Cloud project
Step 2
Request GPU quota
Raise the project's quota for the GPU flavor and region you need
Step 3
Launch an instance
Choose a GPU flavor, region, and image in the Control Panel, the OpenStack API, or Terraform
Step 4
Or call AI Endpoints
Generate an API key and send OpenAI-compatible requests to a hosted model
Where this page's facts come from
- GPU prices
- OVHcloud (ovhcloud.com) · last collected Nov 11, 2026
- LLM API prices
- ovhcloud.com · last collected Nov 11, 2026
- Profile
- Reviewed Sep 28, 2026
- Method
- How we collect and check prices
Frequently Asked Questions
What GPU types does OVHcloud offer?
OVHcloud offers various GPU types including H200 SXM, A10, A10, A10, Quadro RTX 5000, Quadro RTX 5000, Quadro RTX 5000, L40S, L40S, L40S, H100 PCIe, H100 PCIe, H100 PCIe, Tesla V100, Tesla V100, Tesla V100, L4, L4, L4. Check the pricing table above for current availability and pricing.
How do I get started with OVHcloud?
Create an account, Request GPU quota, Launch an instance, Or call AI Endpoints
What are OVHcloud's main advantages?
OVHcloud's main advantages include: European provider with data centers in France, Germany, the UK, and Poland, relevant for GDPR and data-residency requirements, Hourly and monthly billing on the same GPU flavors, Both GPU instances and a per-token inference API from one account, Published list prices for every GPU flavor.
What are OVHcloud's limitations?
OVHcloud's main limitations include: GPU lineup centers on H200, H100, L40S, and L4; no Blackwell instances on the public price list, Most GPU flavors are available in only one or two regions, Part of the GPU list is older V100 and V100S hardware, The A100 form factor (PCIe or SXM) is not stated on the price list.
Compare Providers
Find the best prices for the same GPUs and models from other providers · All OVHcloud alternatives, ranked →
All OVHcloud Comparisons
The weekly price report
One email every Monday with the week's notable GPU and LLM API price moves, plus occasional product news. No spam.