Cerebras and Oracle announce planned OCI inference offering
Cerebras Systems Inc. →Cerebras and Oracle announced plans to make Cerebras inference available through Oracle Cloud Infrastructure for AI laboratories, enterprises and developers. The proposed service extends Oracle's existing infrastructure and inference offerings, including Dedicated AI Clusters in OCI Enterprise AI. The companies described interactive applications and multi-step agents as target workloads, where repeated model calls make latency consequential. Cerebras contributes its Wafer-Scale Engine architecture, which integrates compute and memory to reduce data movement and communication between chips. Availability on OCI was planned for calendar 2027. The announcement describes a planned product integration, not a completed service launch, and explicitly states that development, release timing and functionality remain subject to change.