Local LLM Builder Weighs $6,275 Used WRX80 Workstation vs. More RTX 3090s
infectiousstupidity · reddit · 2026-10-11
A Reddit user running 2×RTX 3090s with vLLM (Qwen3 27B FP8) for coding agents weighs long-term upgrade paths: a used Supermicro WRX80 workstation (Threadripper PRO, 6×PCIe x16, 2000W PSU, bundled RTX 4090) at $6,275 at auction, versus stacking consumer PCs or hunting scarce used EPYC platforms. Includes local used-GPU price data (3090 $1,250, 5060 Ti 16GB $680+, 4090 $2,300) and concerns about buying a platform that ages poorly within 1–2 years.
More from Infra
- Neoclouds that just buy power and GPUs will lose to software-first players, says Beam founder — edgarpavlovsky · 2026-10-11
- M5 Ultra 256GB vs dual DGX Spark: local LLM buyer weighs throughput vs memory — MasterNomie · 2026-10-11
- AI Infrastructure Borrowing Slumps 80% in Three Months, From $113B to $23B — TansuYegen · 2026-10-11
- OpenAI, Microsoft, Google and 4 others signed a pledge to cover AI data centers' grid costs — ChrisGPT · 2026-10-11
- One-town monopolies: Yiwu makes 80% of Christmas decor, every EUV machine comes from Veldhoven — sahilypatel · 2026-10-11
- SemiAnalysis: Anthropic's deal is a four-party trade — 350MW of Rubin capacity at ~$5.80/GPU-hr — AccBalanced · 2026-10-11