Comment by The_Colonel

> enabling consumers to run big-LLM inference locally

A non-technical reason is that the market of people wanting to run their personal LLMs at home is very small.