OpenAI has made a multi-year partnership with Cerebras Systems – an AI company that builds specialized supercomputers and a cloud services platform for deep learning and high-performance computing (HPC). The deal is an investment of over $10 billion to deploy Cerebras’ unique wafer-scale AI computing power globally. This strategic alliance aims to dramatically accelerate AI inference, promising up to 15 times faster responses for OpenAI’s most advanced models.
The deal centers on the Cerebras CS-3 systems, whose single, dinner-plate-sized processors eliminate the bottlenecks of traditional GPU clusters. This ultra-low latency architecture is specifically targeted at powering high-speed, interactive AI applications such as agentic workflows, real-time voice assistants, and coding copilots. With 750 megawatts of Cerebras compute slated for deployment, OpenAI is signaling a major shift in its strategy, prioritizing instantaneous user experience as the next frontier in artificial intelligence.
