Network status Portal Inquiry
Servers and infrastructure

Servers and GPU performance on our own hardware.

One digital partner for computing power — from a single dedicated server to a GPU cluster for model training.

Why only own hardware

We do not host foreign equipment

We do not run colocation — only our own hardware that you rent as a service. Thanks to that we have full control over what runs in the infrastructure, and the same team that deploys a server also manages it.

WHAT IT MEANS FOR YOU
  • Uniform, monitored hardware without unknown variables
  • Faster remote hands — the technicians know every machine
  • You change the configuration remotely through the portal or a quote request
Dedicated servers

Configurator

CPU

AMD EPYC 9354 · 32 cores

1

RAM

DDR5 ECC, 32 GB modules

128 GB

Storage

NVMe SSD, RAID 1

2× 2 TB

Connectivity

From our own network

10 Gb/s
≈ 9 500 Kč/mo.
Send configuration
Server rack cabling detail
GPU compute

Computing power to match your needs

Over 200 GPUs in our infrastructure — from consumer class to datacenter cards with 96 GB of VRAM. We scale the capacity according to what you need to train or infer.

GPU cards in a server

GPU Standard

available
32 GB VRAM16 vCPU64 GB RAM
8,60 Kč/hour
Launch instance

GPU Pro

available
48 GB VRAM24 vCPU96 GB RAM
10 Kč/hour
Launch instance

GPU Ultra

available
96 GB VRAM32 vCPU128 GB RAM
20 Kč/hour
Launch instance

Instances run on our own infrastructure — low latency to our network, no resale through a third party.

Prices are tied to the current market rates of the GPU cloud market plus our margin, according to the USD/CZK rate — we recalculate regularly.

Custom business infrastructure

Infrastructure designed for your operation

For operations that need more than a configurator can do — private VLAN topologies, disaster recovery between locations, a direct connection to GPU compute or dedicated network capacity. We handle these as a project with its own service level agreement.

Send quote request
FAQ

Servers and compute

How quickly is a new server available to me?

We deploy standard configurations quickly — for custom builds it depends on component availability; we will confirm when you send the request.

Can I increase performance later?

Yes, RAM, storage and the GPU class can be upgraded — contact support or use the portal.

How does data backup work?

Basic snapshots are included in the price of the server; extended backup plans are arranged individually according to data volume.

Can I combine GPU compute with a dedicated server?

Yes, both services run on the same infrastructure and can be interconnected over an internal network without the public internet between them.

What happens during a hardware failure?

We have replacement capacity right in the infrastructure — moving you to new hardware is handled by our team, not by you.