The entire card, for your computations
The GPU and its video memory are reserved for you: no MIG partitioning or time-sharing.
WISPHOST / DEDICATED GPUAnonymous offshore
Anonymous offshore
GPU hosting
A fully dedicated NVIDIA GPU for your AI models, rendering and intensive computations. From the RTX A4000 to the H100, choose your resources in six regions, with no KYC and crypto payment.
NVIDIA RTXA100 & H10016 to 80 GB of VRAM
Choose my GPU dedicated physical GPUThe entire card and all of its VRAM.
No KYCOne account, your private access.
Crypto paymentSeven crypto and network choices.
THE RIGHT GPU, IN THE RIGHT PLACE
Have a project in mind? Compare useful power.
Graphics card, video memory, and server resources: everything is there.
Your locationOne chosen region, five up-to-date prices.
Amsterdam · Netherlands · USD · Monthly billing
| Configuration | GPU & video memory | vCPU & RAM | Storage | Network | Monthly price | Configure |
|---|---|---|---|---|---|---|
| 01 / COMPUTEWisp Compute16Entire dedicated card | NVIDIA RTX A400016 GBGDDR6 | 8 vCPU64 GB of RAM | 1 TB NVMeLocal storage | 1 GbpsUnlimited bandwidth | 88.00 $/month | |
| 02 / COMPUTEWisp Compute24Entire dedicated card | NVIDIA RTX 409024 GBGDDR6X | 16 vCPU128 GB of RAM | 2 TB NVMeLocal storage | 2 GbpsUnlimited bandwidth | 188.00 $/month | |
| 03 / COMPUTEWisp Compute32Entire dedicated card | NVIDIA RTX 509032 GBGDDR7 | 24 vCPU192 GB of RAM | 2 TB NVMeLocal storage | 2 GbpsUnlimited bandwidth | 278.00 $/month | |
| 04 / TENSORWisp Tensor A80Entire dedicated card | NVIDIA A10080 GBHBM2e | 32 vCPU256 GB of RAM | 4 TB NVMeLocal storage | 5 GbpsUnlimited bandwidth | 689.00 $/month | |
| 05 / TENSORWisp Tensor H80Entire dedicated card | NVIDIA H10080 GBHBM3 | 48 vCPU384 GB of RAM | 8 TB NVMeLocal storage | 10 GbpsUnlimited bandwidth | 1189.00 $/month |
Dedicated card, without partitioning. IPv4 and IPv6 included. DDoS protection included.
Availability and commissioning to be confirmed with the team.
MORE THAN TERABYTES AND TERAFLOPS
An environment to move forward.
From GPU to data,
the essentials are included.
Your work environment
Start from a bare Linux system or choose a CUDA, PyTorch, or TensorFlow image.
Data within reach of the GPU
The local NVMe storage hosts your datasets, working files, and checkpoints.
Unlimited traffic
From 1 to 10 Gbps depending on the configuration, with unmetered bandwidth.
IPv4 and IPv6
Both protocols are included to connect your environment and your applications.
DDoS protection included
Network protection accompanies every configuration in the GPU catalog.
YOUR SOFTWARE STACK
A foundation suited to your tools.
Choose your image in the configurator.
Bare Linux or compute environment.
Bare system
Install your own software stack on Linux.
CUDA + cuDNN
A foundation for your accelerated compute applications.
PyTorch
A Linux environment for your models and training.
TensorFlow
A Linux image geared toward machine learning.
THE DETAILS THAT MATTERYour next
Your next
compute step.
Understand the resources and prepare your order.
Talk to the teamHow do you choose video memory?
VRAM is the memory of the GPU; it is distinct from the server's RAM. Compare the size of your models, scenes, and processing tasks against the 16, 24, 32, or 80 GB of VRAM offered. The table separately shows the RAM available for the system.
Is the card shared?
Each configuration includes a whole physical GPU, with all of its video memory. It is neither partitioned into MIG instances nor time-shared with another client.
How does regional pricing work?
Choose a region to refresh the five prices. Amsterdam is selected by default. The summary then keeps the region, software image, and monthly amount of your configuration. All prices are in US dollars.
How does the order process work?
Configure your GPU, log in or create your account, then choose your cryptocurrency and its network. Payment instructions are displayed on a separate step. Availability and commissioning must be confirmed with the team.