public cloud compute  gpu

Specialist cloud servers, designed for massively parallel processing

GPU instances employ NVIDIA Tesla V100 graphic processors to deliver the power and performance needed for massively parallel processing, including machine learning and deep learning projects. In line with the OVHcloud ethos, they are available on demand and under an hourly, pay-as-you-go billing method. 

Powered by NVIDIA Tesla V100 GPUs

Designed for use in datacentres, these GPUs are among the most powerful on the market, and have proven invaluable in the fields of artificial intelligence (AI) and graphic computing.


To provide the best possible user experience, OVH and NVIDIA have partnered to offer a best-in-class GPU-accelerated platform, for deep learning and high-performance computing and ​artificial intelligence (AI). It is the simplest way to deploy and maintain GPU-accelerated containers, via a full catalogue. Find out more.

1-4 Tesla cards, with guaranteed performance

Tesla cards are delivered directly to the instance via PCI Passthrough, without a virtualisation layer, so their full power is dedicated to you. As a result, the hardware devotes all of its computing power to your application. Up to four cards can be connected, to further enhance their performance. 

NVIDIA Tesla V100 Features

Performance with NVIDIA GPU Boost

Bidirectional connection bandwidth

CoWoS Stacked HBM2 memory
  • double-precision: 7 teraFLOPS
  • simple-precision: 14 teraFLOPS
  • deep learning: 112 teraFLOPS
  • PCIe 32 GB/s
  • capacity: 16 GB HBM2
  • bandwidth: 900 GB/s


Use cases

Image recognition

Extracting data from images to classify them, identify elements, or build richer documents is now standard practice in many industries. Frameworks like Caffe2, combined with Tesla V100 GPUs, are a key part of this, for projects such as medical imaging, social networks, public protection and security.

Situation analysis

When quick responses to varied and unpredictable situations are required (for self-driving cars and internet network traffic analysis, for example), real-time analysis is essential. This is where deep learning comes in, forming neural networks that learn independently through a training stage.

Human interaction

In the past, people people learned to communicate with machines. In the modern era, machines are learning to communicate with people. Through speech recognition and the identification of emotions from sound and video, tools like Tensorflow are transforming these interactions, opening up a world of new opportunities.

Ready to get started?

Create an account and launch your services in minutes

How to



Launch your instance by choosing a T1 model and NGC image to suit you.



$ docker pull
$ nvidia-docker run nvidia/tensorflow t1



Your AI framework is now ready to go!

tarif public cloud

GPU billing

GPU instances are billed like all of our other instances, on a pay-as-you-go basis at the end of each month. The price depends on the size of the instance you have booted, and the duration of its use.