This is just an uncut wafer - I don't think it's intended to be a wafer-scale chip.
replyCerebras etch memory onto the wafer alongside the processing elements, but AFAIK OpenAI are going to be using HBM memory and a conventional chiplet design.
Still competition for cerebras. Seems quite unlikely they will get an OpenAI deal anytime soon.
replyThey have an OpenAI deal right now. https://openai.com/index/cerebras-partnership/
replyNo - this is OpenAI trying to complete with Google (TPU) and Amazon/Anthropic (Trainium) on cost.
replyCerebras are addressing very specific use cases, not general purpose LLM serving, and OpenAI does already partner with them.