If you had this for your business could this approach be faster than RAG? Input is parsed one token at a time right? Can you cache the state after the initial

If you had this for your business could this approach be faster than RAG? Input is parsed one token at a time right? Can you cache the state after the initial prompt has been provided?