Back to products Platform Products

GPU Server

Our GPU Servers with a choice of multiple Nvidia cards master your AI challenges and are available from CHF 1098.– per month. A complete price list with additional options can be found in the Price Calculator and for more technical details, please check our Documentation.

Explore Go to Cockpit Price Calculator Documentation

The Power You Need for Your AI Tasks

You would like to run a Large Language Model (LLM) that works with human language on the basis of deep learning models? Our GPU servers meet the requirements of high-performance computing – and are also available as a managed option.

Fly to
Highlights Information Case Studies Documents Compatible With Blog posts Price Calculator
Powerful AI Hardware at a Special Rate

Value for Money: Rent One of Our GPU Dedicated Root Servers

You’d like to see the exhilarating GPU server performance with your own eyes and experience it firsthand? Then we have the perfect offer for you: at Nine, you can rent Swiss Dedicated Root Servers equipped with an Nvidia Tesla A100 (40GB) at a special price. Plus, you have the opportunity to test one of these GPU servers free of charge for two weeks.

Contact us now
Full Power
With our GPU servers, you get the computing power AI tasks need nowadays – from running a large language model (LLM) and machine learning to rendering and computer-aided design (CAD), virtualization and simulations. For such complex calculations, GPUs are preferred over CPUs because the numerous processor cores (shaders) can perform countless similar operations in parallel.
State-of-the-art Components
All the other server specifications are equally impressive, such as an Intel Single or Dual Xeon Silver 4210R CPU with 2.4 GHz, 384 GB RAM and two 1.92 TB NVME discs or two 1.92 TB SSDs. Of course, in addition to this special offer, we also offer our cutting-edge servers with H100 or L40s GPUs and AMD Epyc 7763P CPUs as usual.
Rental Prices & Conditions
For just CHF 850.– per month, you can rent one of our Swiss Dedicated Root Servers with a Single Xeon CPU (CHF 1,100.– per month with a Dual Xeon CPU) and experience full GPU power yourself. This product is not listed in our price calculator as it is a special offer and only available while supplies last. To rent such a GPU server, please contact us directly via email at info@nine.ch or by phone at +41 44 637 40 00.
Testing Terms & Conditions
We’ll be handling the requests on a «first come, first served» basis – with a waiting list – and the offer is only available to companies based in Switzerland. To try out a GPU server for two weeks free of charge, please contact us via email at info@nine.ch or by phone at +41 44 637 40 00.
Highlights

Advantages at a Glance

Performance

The computing power of a base system, paired with the power of up to two entry-level or production-level graphics cards

Upgradability

On request, we can equip your GPU server with up to 2 TB RAM, 4x NVME, 8x SATA and two GPUs

Modern hardware

We offer the latest graphics cards and state-of-the-art hardware components for our GPU servers

Managed Server

We also offer GPU servers as a managed variant, where we take on server management for you – we are happy to make you an offer

Swiss location

You can be sure that your data will remain in Switzerland and that your provider follows and embodies a sensitive approach to information security

Hypervisor option

You can also use our base system without a GPU, for virtualisation purposes

Value for money

Starting at CHF 1098.– per month (Root Dedicated Server GPU & Nvidia RTX), you’ll get an up-to-date GPU Server – upgradable to a managed version

Information

About the Product

In contrast to conventional servers, GPU servers offer a drastic increase in performance for complex calculations. This type of server performs the majority of calculations in graphics processors. This is why GPUs are also the most important components for High Performance Computing (HPC). These types of servers are increasingly being used for machine learning or large language models in particular.

More about GPU Server
CPU versus GPU
Nowadays, graphics cards are not only used for image processing, but increasingly in specialised servers for applications such as artificial intelligence (AI), large language models (ChatGPT, Mistral, Jasper, Claude, Ollama, Bard, Gemini etc.), rendering, computer-aided design (CAD), visualizations or simulations. Thanks to the many processor cores (also known as shaders) in GPUs, they can perform many similar operations at the same time. This means that tasks can be processed quickly and cost-effectively.
Nine GPU Server
Our base system comes pre-equipped with powerful components, so you don’t have to make any compromises. The base system can also be used without a GPU as a powerful hypervisor system for virtualisation. Or you can have your Nine GPU Server equipped with one or even two graphics cards – depending on your requirements, you can choose between an entry-level and a production-level model. As a further option, we also offer our Nine GPU servers as managed servers and take care of maintenance, management, monitoring, updates and much more for you – do not hesitate to contact us for an offer.
State-of-the-Art Hardware
When selecting the hardware components, we focussed primarily on an optimal price-benefit ratio in the interest of our customers. We began by examining various options, including the H100 NVL, the L40s and the RTX 4500 ADA from Nvidia, compared their value proposition and availability, and decided on the aforementioned configurations. As flexibility is very important to us at Nine, we can also install further professional components on request. Please note that availability can fluctuate due to the very high demand for GPU servers.
Case Studies

Use Cases for Nine products

Peak Privacy

We supported Peak Privacy in designing and building a specialized server infrastructure for LLM inference. The process was characterized by close collaboration, including intensive idea exchange and joint problem-solving, where the specific expertise of both partners was used optimally to align the hardware precisely with the software requirements.

Peak Privacy

Questions? Our experts are here to help.

Support Contact Sales