{"title":"AI and GPU servers","description":"\u003cp\u003eServers with NVIDIA RTX 4090, RTX 5090 and A100 graphics cards for running language models in your company, fine-tuning models, generating images and video, and rendering. When the server is not busy with your workloads, it can mine cryptocurrencies on its graphics cards.\u003c\/p\u003e\u003cp\u003eWe build every server to order and tailor the configuration to your workload and budget. We agree in advance what will run on it and deliver it with pre-installed software – you switch it on and it works. You can find out how we proceed on the \u003ca href=\"\/en\/pages\/ai-servery\"\u003eAI servers\u003c\/a\u003e page.\u003c\/p\u003e","products":[{"product_id":"ai-server-4x-rtx-5090-128gb-vram","title":"AI server 4× RTX 5090 – 128 GB VRAM, AMD EPYC, 512 GB RAM","description":"\u003ch3\u003eKey features\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e4× NVIDIA GeForce RTX 5090 32 GB GDDR7 (Blackwell) – 128 GB of graphics memory in total\u003c\/li\u003e\n\u003cli\u003eAMD EPYC 7643 processor (48 cores \/ 96 threads) and 512 GB of DDR4 ECC memory\u003c\/li\u003e\n\u003cli\u003e4U rack design, two synchronised 2 kW power supplies, 2× 10 GbE and IPMI remote management\u003c\/li\u003e\n\u003cli\u003eDelivered with pre-installed software for whatever you want to run on the server – it is ready to work as soon as you switch it on\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eConfiguration\u003c\/h3\u003e\u003ctable\u003e\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eGraphics cards\u003c\/td\u003e\n\u003ctd\u003e4× NVIDIA GeForce RTX 5090 32 GB GDDR7 (575 W, PCIe 5.0 x16, running at PCIe 4.0 x16 on this board)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eProcessor\u003c\/td\u003e\n\u003ctd\u003eAMD EPYC 7643 (48C\/96T, 225 W, 128 PCIe 4.0 lanes)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMotherboard\u003c\/td\u003e\n\u003ctd\u003eASRock Rack ROMED8-2T (7× PCIe 4.0 x16, 8× DDR4 ECC, 2× 10 GbE, IPMI)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory\u003c\/td\u003e\n\u003ctd\u003e512 GB DDR4-2666 ECC RDIMM (8× 64 GB, all slots populated)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eStorage\u003c\/td\u003e\n\u003ctd\u003e2 TB NVMe M.2 (PCIe 4.0)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower supply\u003c\/td\u003e\n\u003ctd\u003e2× 2 kW ATX, synchronised (4 kW in total)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eChassis and cooling\u003c\/td\u003e\n\u003ctd\u003e4U rack, tower CPU cooler, 3× 120 mm at the front + 1× 120 mm at the rear\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower draw\u003c\/td\u003e\n\u003ctd\u003eapprox. 2,650 W at full load (cards 4× 575 W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\u003c\/table\u003e\u003ch3\u003eWho it is for\u003c\/h3\u003e\u003cp\u003eFor companies, developers and studios that want to run AI models in-house – your data never leaves your company and you pay no monthly fees for cloud GPUs. 128 GB of graphics memory is enough for language models of up to roughly 120 billion parameters, for image and video generation and for rendering. When the server is not busy with your workloads, it can mine cryptocurrencies on its graphics cards.\u003c\/p\u003e\u003ch3\u003eOperation and installation\u003c\/h3\u003e\u003cp\u003eThe server belongs in a rack or a technical room with good cooling and a circuit that can handle around 2.7 kW. We agree in advance what will run on the server (e.g. vLLM, llama.cpp, Ollama, ComfyUI) and prepare it precisely for your task. We can also advise you on where to place it.\u003c\/p\u003e","brand":"PC Praha","offers":[{"title":"Default Title","offer_id":53063847706911,"sku":"PCP-AI-4X5090","price":1199000.0,"currency_code":"CZK","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0976\/4693\/6351\/files\/k-ai-128-rome-5090-6704tops-4-rtx-5090-blackwell-ai-server.jpg?v=1792300733"},{"product_id":"ai-server-pcpraha-epic-elite","title":"AI server PC Praha Epic Elite – 2× RTX 4090, AMD EPYC 9754, 384 GB RAM","description":"\u003ch3\u003eKey features\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e2× NVIDIA GeForce RTX 4090 24 GB GDDR6X (MSI SUPRIM X) – 48 GB of graphics memory in total\u003c\/li\u003e\n\u003cli\u003eAMD EPYC 9754 processor with 128 cores \/ 256 threads and 384 GB of DDR5 ECC memory\u003c\/li\u003e\n\u003cli\u003e4U rack form factor, ASRock Rack motherboard with 2× 10 GbE and IPMI remote management\u003c\/li\u003e\n\u003cli\u003eSupplied with pre-installed software for whatever you want to run on the server – it is ready to work as soon as you switch it on\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eConfiguration\u003c\/h3\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eComponent\u003c\/th\u003e\n\u003cth\u003eSpecification\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003c\/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eGraphics cards\u003c\/td\u003e\n\u003ctd\u003e2× MSI GeForce RTX 4090 SUPRIM X 24G (24 GB GDDR6X, PCIe 4.0 x16)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eProcessor\u003c\/td\u003e\n\u003ctd\u003eAMD EPYC 9754 (128C\/256T, 360 W, socket SP5)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMotherboard\u003c\/td\u003e\n\u003ctd\u003eASRock Rack GENOAD8X-2T\/BCM (DDR5 ECC, 2× 10 GbE, IPMI)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory\u003c\/td\u003e\n\u003ctd\u003e384 GB DDR5-5600 ECC\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eStorage\u003c\/td\u003e\n\u003ctd\u003eBay for U.2\/U.3 NVMe drives – drive capacity specified in the quote\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower supply\u003c\/td\u003e\n\u003ctd\u003e2,600 W PSU (LX2600W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eChassis and cooling\u003c\/td\u003e\n\u003ctd\u003e4U rack (24″ depth), Supermicro SNK-P0084AP4 CPU cooler\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower draw\u003c\/td\u003e\n\u003ctd\u003eapprox. 1,400 W at full load (cards 2× 450 W, processor 360 W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\n\u003c\/table\u003e\u003ch3\u003eWho it is for\u003c\/h3\u003e\u003cp\u003eFor companies, developers and studios that want to run AI models in-house and also need plenty of CPU cores and memory. 48 GB of graphics memory is enough for language models of up to roughly 30 billion parameters (e.g. Qwen3 32B in 8-bit quantisation), quantised models of around 70 billion parameters with a shorter context, image and video generation, and rendering. 128 cores and 384 GB of memory are well suited to data preparation, virtualisation and CPU workloads. When the server is not busy with your jobs, it can mine cryptocurrencies on its graphics cards.\u003c\/p\u003e\u003ch3\u003ePre-installation, testing and service\u003c\/h3\u003e\u003cp\u003eWe agree in advance what will run on the server (e.g. vLLM, llama.cpp, Ollama, ComfyUI, Blender) and prepare it precisely for your task. Before handover, we test the server for 72 hours under full load. Standard configurations are usually built and tested within 7–14 working days of quote approval. We keep key spare parts in stock in Prague and provide on-site service and part replacement within 24\/48 hours. The warranty is 24 months.\u003c\/p\u003e\u003ch3\u003eOperation and placement\u003c\/h3\u003e\u003cp\u003eThe server belongs in a rack or technical room with good cooling and a circuit that can handle roughly 1.5 kW. Don't want to deal with space, noise and cooling? We can place the server in our hosting at the data centre – terms agreed individually.\u003c\/p\u003e\u003cp\u003e\u003cem\u003ePhoto is for illustration only.\u003c\/em\u003e\u003c\/p\u003e","brand":"PC Praha","offers":[{"title":"Default Title","offer_id":53066123051295,"sku":"PCP-AI-EPIC-2X4090","price":314713.0,"currency_code":"CZK","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0976\/4693\/6351\/files\/kentino-ai-server-4u-rack-front-fans_b66a5ebc-fb18-4142-a6e7-92976bc1345f.jpg?v=1792404297"},{"product_id":"ai-server-and-hpc","title":"AI server HPE ProLiant DL385 Gen10 – 2× NVIDIA A100 80 GB, 2× AMD EPYC 7502","description":"\u003ch3\u003eKey features\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e2× NVIDIA A100 80 GB – 160 GB of HBM2e graphics memory in total, accelerators designed for data centres\u003c\/li\u003e\n\u003cli\u003e2× AMD EPYC 7502 (64 cores \/ 128 threads in total) and 512 GB of DDR4 ECC memory\u003c\/li\u003e\n\u003cli\u003eBranded HPE ProLiant DL385 Gen10 rack server, redundant 2× 1200 W power supply and iLO 5 remote management\u003c\/li\u003e\n\u003cli\u003eEverything new – the server and all components are new (condition: New)\u003c\/li\u003e\n\u003cli\u003eSupplied with pre-installed software for whatever you want to run on the server – it is ready to work as soon as you switch it on\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eConfiguration\u003c\/h3\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eComponent\u003c\/th\u003e\n\u003cth\u003eSpecification\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003c\/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eServer\u003c\/td\u003e\n\u003ctd\u003eHPE ProLiant DL385 Gen10 (CTO), rack-mount\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eGraphics accelerators\u003c\/td\u003e\n\u003ctd\u003e2× NVIDIA A100 80 GB HBM2e\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eProcessors\u003c\/td\u003e\n\u003ctd\u003e2× AMD EPYC 7502 (32C\/64T each, 64C\/128T in total)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory\u003c\/td\u003e\n\u003ctd\u003e512 GB HPE DDR4 ECC\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eDrives\u003c\/td\u003e\n\u003ctd\u003e2× 480 GB M.2 NVMe + 2× 1.6 TB enterprise SAS SSD; backplane for 8× 2.5″ SFF drives\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eStorage controller\u003c\/td\u003e\n\u003ctd\u003eHPE Smart Array S100i (RAID)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eNetwork\u003c\/td\u003e\n\u003ctd\u003e4× 1 GbE + 2× 10 GbE SFP+\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower and management\u003c\/td\u003e\n\u003ctd\u003e2× 1200 W, redundant HPE power supply (1+1), iLO 5 remote management\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCondition\u003c\/td\u003e\n\u003ctd\u003eNew – server and all components new\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower draw\u003c\/td\u003e\n\u003ctd\u003eapprox. 1,200 W at full load (accelerators 2× 300 W, processors 2× 180 W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\n\u003c\/table\u003e\u003ch3\u003eWho it is for\u003c\/h3\u003e\u003cp\u003eFor companies, research teams and developers who need a lot of graphics memory on a single card – for training and fine-tuning models, running large language models and scientific computing. 160 GB of HBM2e memory is enough, for example, for Llama 3.3 70B in 8-bit quantisation with headroom for long context and multiple concurrent users. A100 accelerators can split each card into up to 7 separate instances (MIG), so several teams or jobs can work on the server in isolation. It is also suitable for rendering and big-data analytics.\u003c\/p\u003e\u003ch3\u003ePre-installation, testing and service\u003c\/h3\u003e\u003cp\u003eWe agree in advance what will run on the server (e.g. Ubuntu, NVIDIA drivers, CUDA, Docker, PyTorch, vLLM) and prepare it precisely for your task. Before handover, we test the server for 72 hours under full load. Standard configurations are usually delivered within 7–14 working days of quote approval, depending on component availability. We keep key spare parts in stock in Prague and provide on-site service and part replacement within 24\/48 hours. The warranty is 24 months.\u003c\/p\u003e\u003ch3\u003eOperation and placement\u003c\/h3\u003e\u003cp\u003eThe server belongs in a rack in a server room or data centre. Don't want to deal with space, noise and cooling? We can place the server in our hosting at the data centre – terms agreed individually.\u003c\/p\u003e","brand":"HPE","offers":[{"title":"Default Title","offer_id":53066125443359,"sku":"PCP-AI-HPE-2XA100","price":1234817.0,"currency_code":"CZK","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0976\/4693\/6351\/files\/AI-server-2x-nvidia-tesla-a100-80gb-tensor-core-gpu-server-2x-epyc-7502-cpu-4571-p.jpg?v=1792343619"},{"product_id":"k-ai-96-rome-4090-vykonna-bestie-pro-ai-inferenci-a-llm","title":"AI server K-AI 96 ROME – 4× RTX 4090, 96 GB VRAM, AMD EPYC 7542","description":"\u003ch3\u003eKey features\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e4× NVIDIA GeForce RTX 4090 24 GB GDDR6X – 96 GB of graphics memory in total\u003c\/li\u003e\n\u003cli\u003eAMD EPYC 7542 processor (32 cores \/ 64 threads) and 256 GB of DDR4 ECC memory, expandable to 512 GB\u003c\/li\u003e\n\u003cli\u003e4U rack form factor, two synchronised 2 kW power supplies, 2× 10 GbE and IPMI remote management\u003c\/li\u003e\n\u003cli\u003eMade to order – we build you a new unit from components we have in stock\u003c\/li\u003e\n\u003cli\u003eSupplied with pre-installed software for whatever you want to run on the server – it is ready to work as soon as you switch it on\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eConfiguration\u003c\/h3\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eComponent\u003c\/th\u003e\n\u003cth\u003eSpecification\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003c\/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eGraphics cards\u003c\/td\u003e\n\u003ctd\u003e4× NVIDIA GeForce RTX 4090 24 GB GDDR6X (PCIe 4.0 x16)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eProcessor\u003c\/td\u003e\n\u003ctd\u003eAMD EPYC 7542 (32C\/64T, 225 W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMotherboard\u003c\/td\u003e\n\u003ctd\u003eASRock Rack ROMED8-2T (7× PCIe 4.0 x16, 8× DDR4 ECC, 2× 10 GbE, IPMI)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory\u003c\/td\u003e\n\u003ctd\u003e256 GB DDR4-2666 ECC RDIMM (expandable to 512 GB)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eDrive\u003c\/td\u003e\n\u003ctd\u003e2 TB NVMe M.2 (PCIe 4.0)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower supply\u003c\/td\u003e\n\u003ctd\u003e2× 2 kW ATX, synchronised (4 kW in total)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eChassis and cooling\u003c\/td\u003e\n\u003ctd\u003e4U rack, 120 mm fans with front-to-back airflow\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower draw\u003c\/td\u003e\n\u003ctd\u003eapprox. 2,200 W at full load (cards 4× 450 W, processor 225 W)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eAvailability\u003c\/td\u003e\n\u003ctd\u003eMade to order – we build you a new unit (components in stock)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\n\u003c\/table\u003e\u003ch3\u003eWho it is for\u003c\/h3\u003e\u003cp\u003eFor companies that want to run language models in-house – an internal chatbot, an assistant or working with company documents (RAG). Your data never leaves your company and you pay no monthly fees for cloud GPUs. 96 GB of graphics memory is enough for models of around 70 billion parameters in quantised form (e.g. Llama 3.3 70B), several smaller models at once, fine-tuning (LoRA\/QLoRA) of 7–34-billion-parameter models, and image and video generation (FLUX.1, SDXL). In our test, Llama 3.3 70B (AWQ, 4-bit) reached up to 179 tokens per second with 32 concurrent requests. When the server is not busy with your jobs, it can mine cryptocurrencies on its graphics cards.\u003c\/p\u003e\u003ch3\u003ePre-installation, testing and service\u003c\/h3\u003e\u003cp\u003eWe agree in advance what will run on the server (e.g. Ubuntu, CUDA, Docker, vLLM, ComfyUI) and prepare it precisely for your task. Before handover, we test the server for 72 hours under full load. We have the components in stock – we build a new server from them and test it, usually within 7–14 working days of quote approval. We keep key spare parts in stock in Prague and provide on-site service and part replacement within 24\/48 hours. The warranty is 24 months. We can adjust memory, drives or network cards to suit your needs.\u003c\/p\u003e\u003ch3\u003eOperation and placement\u003c\/h3\u003e\u003cp\u003eThe server belongs in a rack or technical room with good cooling and a circuit that can handle roughly 2.3 kW. Don't want to deal with space, noise and cooling? We can place the server in our hosting at the data centre – terms agreed individually.\u003c\/p\u003e","brand":"PC Praha","offers":[{"title":"Default Title","offer_id":53066125476127,"sku":"PCP-AI-KAI96-4X4090","price":543886.0,"currency_code":"CZK","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0976\/4693\/6351\/files\/k-ai-96-rome-4090-vykonna-bestie-pro-ai-inferenci-a-llm.jpg?v=1792343619"}],"url":"https:\/\/pcpraha.cz\/en\/collections\/ai-servery.oembed","provider":"PCPRAHA","version":"1.0","type":"link"}