Latest
AI Affairs, home

Monday 9 November 2026

Markets · Brief

Google introduces two TPU variations for security and AI agent workloads

The new chips are intended to support confidential computing and give models access to more information when responding to queries.

Google sign outside its data center in The Dalles, Oregon
Photo: Lambtron, CC BY-SA 4.0, via Wikimedia Commons (cropped)

Bloomberg reported on 9 November that Google is introducing two variations of its tensor processing units (TPUs), expanding its challenge to Nvidia in an AI chip market worth hundreds of billions of dollars.

Google’s TPUs are custom processors used to develop and run AI software. The new variations are intended to address confidential computing needs and help models draw on a wider range of information when answering queries, Bloomberg reported. They are also designed for workloads associated with AI agents.

Oracle plans to add Google TPUs to its cloud services, giving customers a processor option alongside Nvidia hardware, News18 reported on 29 October. Oracle also plans to offer processors from Cerebras Systems and SambaNova Systems.

Oracle plans to offer Google TPUs on its cloud platform in 2027, as AI Affairs reported.

Sources

Topics: Chips, Enterprise adoption

What should we cover? Tell us