Oracle teams up with Cerebras for a planned 2027 cloud inference service
The partnership targets the repeated model calls behind multistep AI agents, but it is not a delivery commitment.
Loading page…
The partnership targets the repeated model calls behind multistep AI agents, but it is not a delivery commitment.
Listen to this story
Oracle’s planned Cerebras-powered inference service would expand the compute options available through OCI for latency-sensitive AI applications, including agents that make repeated model calls. Cerebras says its wafer-scale design reduces data movement between chips, but the companies have not published performance measurements for the planned deployment. The 2027 availability remains a roadmap target, not a firm delivery or purchasing commitment, so customers cannot yet assess its real-world speed or cost.
The planned service targets AI labs, enterprises, and developers building interactive applications and multistep agents.
Cerebras says its Wafer-Scale Engine combines compute and memory on a single wafer-scale chip to reduce communication between chips.
Oracle already offers Dedicated AI Clusters through OCI Enterprise AI; Cerebras is intended to add another compute option.
Oracle is lining up another way to run AI applications that need quick responses. On October 27, 2026, it announced a partnership with Cerebras to bring the chipmaker’s inference technology to Oracle Cloud Infrastructure. Availability is planned for calendar year 2027, targeting AI labs, enterprises and developers building interactive applications and agents that carry out multistep tasks.
Inference is the work of running an AI model to produce a response. The companies’ pitch centers on the delays that accumulate across a multistep workflow: each additional model call adds waiting time.
Cerebras addresses that delay with its Wafer-Scale Engine, a processor that integrates compute and memory on a single wafer-scale chip. The company says this arrangement reduces data movement and communication between chips. Its stated benefit is rapid token generation—the production of pieces of an AI response—during long reasoning sequences and multistep agent workflows. The announcement does not include performance measurements for the planned OCI deployment.
Every step an agent makes requires inference, whether it’s reasoning, checking its work, or taking action. The faster the inference, the more work AI can do without making users wait.
Andrew Feldman, Cerebras co-founder and CEO
For Oracle, Cerebras would join an existing set of infrastructure and inference capabilities, including Dedicated AI Clusters in OCI Enterprise AI.
Mahesh Thiagarajan, president of Oracle Cloud Infrastructure, framed the choice around matching compute to an application’s performance needs. He described the planned Cerebras offering as another ultrafast and deterministic option.
The release also carries an explicit purchasing caveat. Oracle and Cerebras say it outlines their general product direction and is not a commitment to deliver particular functionality. Development, release, timing and pricing may change at the respective companies’ discretion. Customers are specifically told not to rely on the announcement when making purchasing decisions.
Loading discussion...
Join the conversation
Explain which would matter more in a task you care about.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.