Optimizing a low latency server for chat AI requires a network first approach where server location and the quality