Oracle announced a partnership with Cerebras Systems to integrate wafer-scale inference technology into the Oracle Cloud Infrastructure (OCI). The agreement was officially disclosed on October 27, 2026, during the Oracle AI World event held in Las Vegas.
The integration aims to provide high-speed, deterministic inference options for complex, multi-step AI workflows, such as those required for AI agents. This offering will complement Oracle's existing portfolio, which currently includes Dedicated AI Clusters.
Cerebras utilizes its Wafer-Scale Engine, which combines compute and memory on a single chip to achieve ultra-low latency. The service is intended for AI labs, enterprises, and developers looking to accelerate agent-based tasks.
The companies have not yet disclosed details regarding pricing, capacity, or the scale of deployment. Access to the new inference capabilities is expected to become available to customers in 2027.
