I find it hard to see how that would ever be economical. LLMs need very expensive power hungry chips and datacenters have - bulk discounts - cheaper electricit

I find it hard to see how that would ever be economical. LLMs need very expensive power hungry chips and datacenters have - bulk discounts - cheaper electricity - high utilisation to spread the costs among many users I don't see how PCs could ever compete against it. Most users AI demands would probably result in >90% idle time on the GPU.