What are vGPU servers and for which scenarios are the new products suitable?

vGPU servers are powerful root servers with permanently assigned, virtualized GPU performance (vGPU). They are particularly suitable for compute-intensive applications such as AI inferencing, video processing, 3D visualizations and data-driven analyses.

At which locations are the vGPU products from netcup available?

The vGPU servers are hosted in our highly secure data center in Nuremberg (Germany) - fully GDPR-compliant and with short latency paths within Europe .

Are the netcup vGPU products suitable for AI training?

Our vGPU servers are designed for AI inferencing - i.e. for the productive use of pre-trained models . They are not intended for large-scale AI training.

Do I need a virtual server (VPS) instead of a vGPU?

If you don't need a GPU and CPU/RAM are sufficient, virtual servers (VPS) are a cost-effective choice – flexible and with hourly billing.

Do the vGPU products support CUDA Technology?

Yes, the NVIDIA™ H200 GPUs used support the CUDA platform and thus enable GPU-accelerated computing processes and parallel data processing.

Does the best price guarantee and satisfaction guarantee apply to these products?

The best price and satisfaction guarantee does not apply to these products.

Which images are used to deploy the vGPU servers?

The vGPU servers are currently provided exclusively with an Ubuntu image . This is because the required NVIDIA drivers are not freely available and are officially supported primarily on Ubuntu. As a result, no alternative images are offered at this time, and the vGPU servers can only be operated with the supplied Ubuntu image.

When are ARM servers an alternative to vGPU?

If your stack is ARM64-compatible and primarily CPU-bound, ARM servers offer an energy-efficient option with no GPU overhead.

I would like to use the new servers for LLMs. Which language models can be operated on the servers with the new tariffs?

Our vGPU servers are ideal for more compact language models such as Llama3, Mistral, Gemma, Qwen2 or Phi3 - ideal for high-performance and resource-saving inference.

How are the servers billed?

Billing is monthly , based on a minimum contract term of just one month. So you remain flexible with full cost control.

Which hypervisor is used for instance virtualization?

As with all our root servers, the vGPU instances also use the KVM hypervisor in the Linux kernel - for stable and high-performance virtualization.

What other server options does netcup offer?

All server variants at a glance: VPS, root servers, ARM servers, and vGPUs, including local block storage extensions.

My question is not listed here. Where can I find an answer?

Our support team is always available to answer any questions you may have. Click here for the support form.

vGPU server

High-performance vGPU power from the cloud - available quickly and cost-effectively
High-end hardware (NVIDIA™ H200, AMD EPYC™ 9535)
Minimum contract term of just one month
Guaranteed vCPU and VRAM resources
Hosted in Europe / Server location Nuremberg

GPU performance from the cloud for demanding workloads

Powerful root servers with dedicated vGPU performance - perfect for compute-intensive tasks and a real alternative to physical GPU hardware.

RS 2000 vGPU 7

NVIDIA™ H200
7 GB VRAM (dedicated)
AMD EPYC™ 9535
8 vCore (dedicated)
16 GB DDR5 RAM (ECC)
512 GB NVMe SSD
Location: Nuremberg, Germany

already from monthly

€82.31

(incl. 0% VAT)

Order now

RS 4000 vGPU 14

NVIDIA™ H200
14 GB VRAM (dedicated)
AMD EPYC™ 9535
12 vCore (dedicated)
32 GB DDR5 RAM (ECC)
1 TB NVMe SSD
Location: Nuremberg, Germany

already from monthly

€157.10

(incl. 0% VAT)

Order now

Click here for the vGPU server FAQs

Our vGPU products are ideal for the following scenarios:

3D rendering & visualizations
AI Inferencing
Video processing / video encoding
Big Data & Analytics / Data Analysis

Our vGPU products offer high-performance GPU power from the cloud - ideal for demanding workloads. With two different tariffs, we specifically cover different requirements so that you get exactly the performance you need. Thanks to dedicated vGPU allocation per virtual machine, you benefit from guaranteed resources and high performance. At the same time, our solutions are a cost-effective alternative to physical GPU hardware.

Your advantages with netcup vGPU

GPU power for demanding workloads

Our root servers with additional GPU power take your compute-intensive applications to a new level. Whether rendering tasks, CAD or AI inferencing - accelerated workloads mean faster results and greater efficiency.

Graphics-intensive applications from the cloud

Use powerful vGPUs for professional 3D designs, visualizations, simulations or video production - without any hardware of your own. Maximum performance, directly from the cloud - flexible, scalable and ready for immediate use.

Secure & isolated - dedicated vGPUs per VM

Thanks to virtualization with dedicated vGPU allocation per virtual machine, we guarantee maximum performance and high security. No resource sharing problems, no compromises - full control for every workload.

Cost-efficient alternative to physical GPU hardware

Save yourself expensive acquisition costs and long set-up times. With our vGPU servers, you get exactly the performance you need - at a fraction of the cost of traditional hardware solutions. Ideal for AI inference, video encoding or data-driven analyses.

Stay in control of your data

As an ISO-certified company, netcup is subject to strict data protection guidelines. The data is hosted in our certified data centers in compliance with the GDPR and is therefore subject to the strictest GDPR-compliant data protection guidelines and security measures.

Local Block Storage for vGPU

You get up to 4 TB of storage with Local Block Storage technology. Choose your desired storage capacity without any contracted term and easily and quickly meet your needs for new resources. Benefit from the flexible and cost-effective option of expanding storage capacity as needed at any time.

Local Block Storage Price Calculator

Volume

Unit

€0.01per month

Price per month from

€0.01

More details

1 TB = 1024 GB / Available for root servers from G9, VPS x86 from G10 and VPS ARM64 G11 (up to max. 8 TB) as well as selected offers of the VPS x86 G11s (up to max. 2 TB). vGPU plans can be extended with up to 4 TB in sum. The price per GB is €0.0119. The total amount is rounded to the nearest cent.

By using modern high-end hardware - including NVIDIA™ H200 GPUs and AMD EPYC™ 9535 processors - our vGPU servers provide an ideal foundation for graphics-intensive and compute-intensive applications.

Whether 3D rendering, video encoding or AI inferencing: With guaranteed vCPU and VRAM resources, you get reliable performance - hosted in our certified data center in Nuremberg and fully GDPR-compliant.