Oracle Brings Google TPUs to Oracle Cloud Infrastructure

Summary

• Oracle and Google announced that Google Tensor Processing Units (TPUs) will be available through Oracle Cloud Infrastructure (OCI).
• This partnership marks the first time Google TPUs are offered on a hyperscale cloud platform outside of Google.
• OCI will provide power, cooling, and datacenter networking to deploy Google TPUs at scale.
• Google TPUs are custom accelerators co-designed with open software to support the entire AI lifecycle, including frontier training, large-scale inference, and agentic AI.
• The TPU architecture is designed to maximize AI processing per watt of electricity with fast access to model data and rapid chip communication.
• Google's eighth-generation TPU portfolio includes TPU 8t for training and TPU 8i for inference.
• TPU 8t delivers nearly 3x the compute performance per pod over the previous generation and cuts frontier model development cycles from months to weeks.
• TPU 8i is designed to handle the iterative work of many specialized agents.
• Google TPUs on OCI will be available in calendar year 2027, ramping over a multi-year period.
• Customers can choose Google TPUs within OCI to optimize for workloads, performance requirements, and cost objectives.