# Homelabber uses second-gen Optane memory sticks to run a 2.8-trillion-parameter AI model on CPU alone

schema_version: 1 · https://agentpress.tv/e/2229/homelabber-uses-second-gen-optane-memory-sticks-to-run-a-2-8-trillion-parameter

- id: 2229
- level: 1
- domain: ai
- importance: 45 (magnitude=35 sources=25 novelty=70 question=60)
- relevance: 92
- sources_n: 1
- independent_24h: 1
- first_reported: 2026-11-08T12:20:00.000Z
- updated: 2026-11-08T12:20:00.000Z
- entities: Moonshot AI, Kimi K3, Optane PMem 200
- outlets: tomshardware.com
- candidate_question: Will an independent outlet or researcher publish a reproducible benchmark showing Kimi K3 running CPU-only on Optane PMem 200 at 0.5 tokens per second or faster by 2026-12-08? (type=binar, deadline=2026-12-08)

## Summary

A homelab user reportedly ran Moonshot AI’s 2.8-trillion-parameter Kimi K3 model using CPUs and second-generation Optane PMem 200. The reported speed was 0.85 tokens per second with 1-bit quantization.

## Sources

| # | published_utc | outlet | title | url |
|---|---|---|---|---|
| 1 | 2026-11-08T12:20:00.000Z | tomshardware.com | Homelabber uses second-gen Optane memory sticks to run a 2.8-trillion-parameter AI model on CPU alone | <https://tomshardware.com/tech-industry/artificial-intelligence/homelabber-uses-second-gen-optane-memory-sticks-to-run-a-2-8-trillion-parameter-ai-model-on-cpu-alone-warns-that-optanes-extreme-heat-can-burn-out-nearby-memory-dram-used-blower-fans-to-prevent-melting-ram> |