Custom AI servers · Prague
AI servers with tailored pre-installation
GPU servers for training and running AI models. We agree in advance what you'll do on the server, design the build around your workload and budget, and deliver it with pre-installed software – switch it on and it works.
We test every server under full load for 72 hours before handover. We deliver it to you or place it in our data centre.
What you'll use an AI server for
We'll design the hardware based on what you'll be doing on it.
- Running language models (LLMs)Most popular Your own chatbot, assistant or company document processing on your own hardware. Your data stays with you and you don't pay per cloud request.
- Training and fine-tuning models Fine-tuning language models on your data, training your own neural networks and computer vision.
- Images, video and rendering Image and video generation (e.g. ComfyUI), rendering in Blender, editing and effects in DaVinci Resolve.
- Development and testing An environment for developers and data analysts – Linux, NVIDIA drivers, CUDA and Docker set up so you can get straight to work.
Why an AI server from us
- Switch it on and it worksWe agree in advance what the server will do and install the software for your workload for you. You don't need to install or tune anything.
- Made-to-order configurationThe builds on offer are a starting point. We'll adapt the graphics cards, memory and drives to your workload and budget.
- 72 hours under loadEvery server goes through a 72-hour stress test before handover. We check stability under full load on the graphics cards and processor, temperatures, and for ECC memory also that it's error-free.
- Service within 24/48 hoursWe keep key spare parts in stock in Prague. We provide service call-outs and parts replacement within 24/48 hours.
- At your premises or in our hostingWe deliver the server to your office or your data centre. Don't want to deal with space, noise and cooling? We'll place it in our data centre.
- We won't leave you on your ownWe take an individual approach to every customer – from the first consultation to operation. If you're ever unsure, just get in touch.
How it works
- ConsultationWe discuss what you want to run on the server – models, software, data volume – where the server will be located and what your budget is.
- Configuration proposalWe design a tailored build: graphics cards, processor, memory, drives and cooling. We send you a no-obligation quote.
- Assembly and pre-installationWe build the server, install the system, drivers and software for your workload, and put it through a 72-hour stress test.
- Handover or hostingWe hand the server over ready for work, or place it in our data centre. We're there for you afterwards too.
Configurations
Sample builds. We'll adapt each one to your workload and budget.
-
Made to order
AI server K-AI 96 ROME – 4× RTX 4090, 96 GB VRAM, AMD EPYC 7542
4× RTX 409096 GB VRAM
543 886 Kč
449 493 Kč excl. VAT
Details and configuration →
-
Made to order
AI server HPE ProLiant DL385 Gen10 – 2× NVIDIA A100 80 GB, 2× AMD EPYC 7502
2× A100 80 GB
1 234 817 Kč
1 020 510 Kč excl. VAT
Details and configuration →
-
Made to order
AI server PC Praha Epic Elite – 2× RTX 4090, AMD EPYC 9754, 384 GB RAM
2× RTX 409048 GB VRAM
314 713 Kč
260 093 Kč excl. VAT
Details and configuration →
-
Made to order
AI server 4× RTX 5090 – 128 GB VRAM, AMD EPYC, 512 GB RAM
4× RTX 5090128 GB VRAM
1 199 000 Kč
990 909 Kč excl. VAT
Details and configuration →
Looking for a different build? Tell us what you need – we'll design a configuration to suit you.
Enquire about an AI server
Tell us what you want to run on the server, where it will run and what budget you have in mind. We'll get in touch, go through the details and prepare a no-obligation quote.
Or write to obchod@pcpraha.cz
AI server FAQ
What does "tailored pre-installation" mean?
Before building the server we agree what you'll use it for – which model, framework or program. Based on that we install the system, drivers and software and test the server. Once switched on it's ready to work; you don't need to set anything else up.
Can I adjust the configuration?
Yes. We build every server to order, and the builds on offer are just a starting point. We'll adapt the graphics cards, processor, memory and drives to your workload and budget.
Do you test the server before handover?
Yes. Every server goes through a 72-hour stress test before handover – we check stability under full load on the graphics cards and processor, temperatures, and for ECC memory also that it's error-free. Only then do we hand it over or place it in our hosting.
Which model will run on the server?
It depends mainly on the graphics memory (VRAM) and the number of cards. Tell us which model or program you want to run and we'll design a build that will run it.
Can the server run in your data centre?
Yes. We can place the server in our data centre hosting – you don't need to deal with space, noise, cooling or power supply. We agree the terms and price of placement individually based on the configuration.
How long does delivery take?
We usually build and test standard configurations within 7–14 working days of the quote being approved. It depends on the availability of specific components – we'll confirm the exact date in the quote.
Will you help after handover too?
Yes, we stay with our customers after handover. We keep key spare parts in stock in Prague and provide service call-outs and parts replacement within 24/48 hours. For businesses we can agree an individual service level agreement (SLA). Repairs are handled in our own service centre in Prague 10.
Can the server also mine?
Yes. When the server isn't running your workloads, it can mine cryptocurrencies on its graphics cards. For cards where we have current network data, you'll find an estimate of today's earnings on the server's page.
Do you sell to businesses?
Yes, we invoice businesses with a company ID and VAT number. For configurations we show prices both including and excluding VAT.
Made to order
Made to order
Made to order
Made to order