OpenAI Signs $10 Billion Deal With Cerebras

OpenAI and Cerebras announced a multi-year compute partnership on 14th January, in which Cerebras will provide dedicated infrastructure for OpenAI's models. 

The arrangement involves a $10 billion commitment from OpenAI to secure computing capacity for running inference at scale. Cerebras stated that the partnership will help OpenAI models deliver faster response times for complex, time-consuming tasks.

OpenAI is no longer relying solely on internally-managed infrastructure or cloud providers like Azure. Instead, it's securing specialized hardware from a company that built processors explicitly designed for AI workloads. Cerebras' WSE (Wafer Scale Engine) chips are built to handle large model inference with lower latency and higher efficiency than traditional GPU architectures.