{"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\/products\/k-ai-96-rome-4090-vykonna-bestie-pro-ai-inferenci-a-llm","provider":"PCPRAHA","version":"1.0","type":"link"}