AI, rendering and computing
AI server for home – local AI on your desk
Want to run language models, generate images or develop AI applications at home or in a small office – without the cloud and without monthly fees? Choose from AI computers and personal AI supercomputers that fit on your desk: AI mini PCs with the NVIDIA GB10 chip (NVIDIA DGX Spark and its variants from ASUS, Dell, GIGABYTE, Lenovo and MSI) and workstations with an RTX graphics card.
Which models fit: an AI mini PC with 128 GB of unified memory can handle gpt-oss-120b or Llama 3.3 70B in 8-bit, and with FP4 quantisation models of up to roughly 200 billion parameters. Depending on memory size, computers with a single card and 16–32 GB of graphics memory can handle quantised models of roughly 8 to 32 billion parameters, image generation and development.
Who it is for: developers, data scientists, students and enthusiasts who want local AI at home with their data under their own control. We explain how to choose in the article AI server for business, work and home; larger builds and how we work together are described on the AI servers page.
- 11 systemsfrom 95 490 Kč incl. VAT
- Pre-installedswitch it on and the server is ready to work
- Made to orderwe adjust the configuration to your needs
Server comparison
Prices excl. VAT, with VAT below. We will adjust the configuration to suit you.
| Server | GPU | VRAM | Processor | RAM | Power draw | Price excl. VAT |
|---|---|---|---|---|---|---|
| Dell Pro Max Tower T2 | 1× RTX 2000 Ada | 16 GB | Intel Core Ultra 7 265 | 32 GB DDR5 | — | 78 917 Kč95 490 Kč incl. VAT |
| HP Z2 G1i | 1× RTX PRO 2000 Blackwell | 16 GB | Intel Core Ultra 9 285K | 64 GB DDR5 | — | 92 388 Kč111 790 Kč incl. VAT |
| HP Z2 G1i | 1× RTX PRO 4000 Blackwell | 24 GB | Intel Core Ultra 9 285K | 32 GB DDR5 | — | 100 983 Kč122 190 Kč incl. VAT |
| AI workstation PC Praha | 1 GPUs of your choice | 16–96 GB | AMD Ryzen 9 9950X | 128–192 GB | — | from 135 860 Kč164 390 Kč incl. VATmade to order |
| ASUS Ascent GX10 | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 137 926 Kč166 890 Kč incl. VAT |
| MSI EdgeXpert MS-C931 | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 141 231 Kč170 890 Kč incl. VATmade to order |
| Lenovo ThinkStation PGX | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 149 496 Kč180 890 Kč incl. VAT |
| GIGABYTE AI TOP ATOM | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 149 496 Kč180 890 Kč incl. VAT |
| ASUS ROG G700 | 1× RTX 5090 | 32 GB | Intel Core Ultra 7 265KF | 64 GB DDR5 | — | 166 107 Kč200 990 Kč incl. VATmade to order |
| NVIDIA DGX Spark | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 182 719 Kč221 090 Kč incl. VAT |
| Dell Pro Max with GB10 | 1× GB10 Grace Blackwell | — | NVIDIA GB10 Grace Blackwell (20 jader ARM) | 128 GB LPDDR5x (unified) | — | 191 066 Kč231 190 Kč incl. VAT |
VRAM = total graphics memory of all cards. Power draw at full load is approximate.
What the servers are for
- AI and LLMs
Language models in your company
An internal chatbot, assistant or working with company documents (RAG). Your data never leaves the company and you pay no monthly fees for cloud GPUs.
- Images, video, 3D
Generation and rendering
Image and video generation (FLUX.1, SDXL, ComfyUI), rendering in Blender and workloads that need lots of cores and memory.
- Bonus
Mining in idle time
When the server isn't running your workloads, it can mine cryptocurrencies on its graphics cards. See how much the cards earn on our graphics card profitability page.
AI server FAQ
Which AI server should I choose for home?
For language models, the most important factor is how much memory the graphics chip has available. An AI mini PC with NVIDIA GB10 has 128 GB of unified memory and is suited to large models and AI application development. A workstation with a single RTX card is more versatile – it handles smaller models, image generation, rendering and everyday work in Windows.
How does the NVIDIA DGX Spark differ from the ASUS Ascent GX10, Dell Pro Max with GB10 and others?
They are all built on the same NVIDIA GB10 Grace Blackwell platform with 128 GB of unified memory and DGX OS. They differ mainly in storage capacity (1 or 4 TB), chassis design and cooling, and the manufacturer's warranty terms. Their performance in AI workloads is very similar.
How large a language model can I run on an AI mini PC?
With 128 GB of unified memory you can run, for example, gpt-oss-120b or Llama 3.3 70B in 8-bit, and with FP4 quantisation models of up to roughly 200 billion parameters. Two units linked via ConnectX-7 can handle models of up to around 405 billion parameters. For fast serving of many users at once, a server with graphics cards is a better fit.
Do I need a special socket or room for an AI mini PC?
No. An AI mini PC with GB10 is powered by an external 240 W adapter, so you plug it into an ordinary socket and put it on your desk. Large rack servers, on the other hand, belong in a technical room or in our hosting.
Will you help with installing the software?
Yes. We advise on choosing a model and on installing and configuring the software – drivers, CUDA, Ollama, vLLM or ComfyUI – as well as connecting it to your home or company network.
Discuss a configuration
Tell us which models or programs you want to run, how many users and what your budget is. We will propose a configuration and send you a quote.
- Phone+420 792 480 835
- Emailobchod@pcpraha.cz
- Shop and service centreRostovská 260/2b, Praha 10 · Mon–Fri 9–17