CNBC reports Cerebras hardware will run on Oracle Cloud Infrastructure

The same expansion covers Google TPUs and the addition of ChatGPT and Codex to Oracle's applications, CNBC reported.

CNBC described the move as part of a wider expansion of Oracle's cloud chip lineup, which also covers Google TPUs and the integration of ChatGPT and Codex into Oracle applications.

A Cerebras spokesperson told CNBC that the Oracle deal is material to the company. CNBC reported that the announcement comes seven months after Amazon Web Services said it would bring Cerebras hardware into its data centers.

Oracle had already named Cerebras in March. CNBC reported on March 10 that Oracle listed Cerebras and Positron as emerging accelerator designs it offers, alongside the latest Nvidia and AMD options.

Other buyers have moved as well. On October 2, the AI cloud startup General Compute signed a multi-year agreement to deploy Cerebras wafer-scale hardware, with access through its cloud platform starting in the first quarter of 2027, DatacenterDynamics reported. In August, Cerebras reported more than 600 megawatts of data center capacity live or under contract for delivery by the end of 2027, and said it was raising manufacturing capacity more than tenfold during 2026, TechCrunch reported.

Oracle plans to use Cerebras equipment in its AI data centers. For investors with money in Oracle, that means a named non-Nvidia supplier is now part of the hardware behind Oracle's AI buildout.