Oracle Lines Up Nvidia, Google and Cerebras AI Chips for Its Cloud

Oracle will expand its cloud AI infrastructure to multi-gigawatt scale with Nvidia Vera Rubin NVL72 systems, and will run Google Tensor Processing Units outside Google's own cloud for the first time. The company announced the deals with Nvidia, Google and Cerebras Systems on Wednesday. Financial terms were not disclosed.

The moves give Oracle Cloud Infrastructure (OCI) customers three new AI compute options — Nvidia GPUs and CPUs for training and agent workloads, Cerebras' fast inference, and Google TPUs — under a single cloud platform.

What does the Nvidia expansion deliver?

Oracle and Nvidia plan to scale OCI Superclusters built on Nvidia Vera Rubin NVL72 systems to multi-gigawatt capacity, responding to customer demand for AI training, inference and agentic applications. Nvidia Vera Rubin NVL72 systems and Nvidia Vera CPUs are now generally available on OCI, serving both GPU-intensive training and CPU-intensive agent tasks.

Nvidia CEO Jensen Huang has previously said Vera Rubin systems can generate up to $40 billion in revenue per gigawatt for the chipmaker.

OpenAI engineers endorsed the platform. Sreekanth Pothanis, engineering leader for compute infrastructure at OpenAI, said Nvidia Vera Rubin on OCI "provides the computing performance and scale needed to accelerate model training." Uday Ruddarraju, chief technology officer of compute at OpenAI, said Nvidia Vera CPUs on OCI "deliver world-class loaded performance for agentic sandbox workloads."

Ian Buck, vice president of hyperscale and high-performance computing at Nvidia, said the companies are bringing Vera Rubin and Vera CPUs to OCI so customers can "put more of their infrastructure to work today."

When do Cerebras and Google TPUs arrive?

Cerebras Inference on OCI is planned for calendar year 2027, offering AI labs, enterprises and developers an inference option where speed is critical. Andrew Feldman, co-founder and CEO of Cerebras, said faster inference "enables developers to build agents that do more, faster."

Google TPUs on OCI will also become available in calendar year 2027 and will ramp up over a multi-year period to meet customer needs. This marks the first time Google's TPUs run on a hyperscale cloud platform outside Google.

Amin Vahdat, senior vice president and chief technologist for AI & Infrastructure at Google, said bringing TPUs to OCI "will give more organisations access to this technology."

How does Oracle position the deals?

Mahesh Thiagarajan, president of Oracle Cloud Infrastructure, framed the three partnerships as one strategy:

• The Nvidia collaboration extends OCI's capabilities across training, inference and agentic AI.
• Cerebras inference adds an "ultrafast and deterministic" option to Oracle's compute portfolio.
• Google TPUs give customers "access to multiple leading AI compute architectures in one cloud."

The announcements position OCI as a multi-vendor AI compute platform: Nvidia silicon for training and agents now, with Cerebras inference and Google TPUs arriving in 2027. One cloud, three competing chip architectures, and a multi-gigawatt buildout to hold them.