{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicfpveuveroeoljpav56xtpnowyo6illxxsn7zb4yhwpajb3mqtdq",
    "uri": "at://did:plc:ax37ypwwfpozwkc552d7scv5/app.bsky.feed.post/3mmk4ywrigel2"
  },
  "path": "/threads/looking-for-suggestions-for-a-llm-dedicated-server-build.2634741/",
  "publishedAt": "2026-05-23T17:59:14.000Z",
  "site": "https://forums.anandtech.com",
  "tags": [
    "Computer Building"
  ],
  "textContent": "Hello.\n\nI was assigned to set up a LLM server with a budget of 14000 USD aprox.\n\nI've built gaming builds before but I'm completely new in terms of Large Language Model. I just started the process of reading all the information and videos available on the internet about this, and I thought I would ask for advice here to see if I can start getting suggestions while I gather more information.\n\nHow much storage space would we need for the models and other data?\n\nIs it better to have two 4090's or four 3090's as the GPU's? Any other suggestions on GPUs for this budget?\n\nHow much RAM would we need?\n\nFor the motherboard and CPU I was thinking of:\nMotherboard: Gigabyte B650 EAGLE AX\nCPU: AMD Ryzen 5 9600X\n\nI'm also thinking on building this on a GPU rack since it will be placed in an open space with little to no dust. Does that sound good?\n\nAny other tips of matters that I might not have thought of?",
  "title": "Looking for suggestions for a LLM dedicated server build?"
}