# We got custom inference per model per hardware now. How long before we get custom hardware for model? — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/4d5dcf154dffc794

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-11-10 22:04 ingested 2026-11-11 02:42

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1xt059z/we_got_custom_inference_per_model_per_hardware/)

## retained text



```
Is it possible to have a custom hardware designed (by ai) and open sourced that we can send for manufacturing for specific models? like for DSF 4.1F 
```



document fetched 2026-11-11 00:42 · canonical https://reddit.com/r/LocalLLaMA/comments/1xt059z/we_got_custom_inference_per_model_per_hardware
