Google Cloud Next 26 Unveils Gemini Enterprise Agent Platform and New TPUs
Google Cloud Next '26 focuses on the agentic enterprise with the introduction of the Gemini Enterprise Agent Platform, replacing the standalone Vertex AI roadmap. The event also features new eighth-gen TPU chips, AI networking fabric, and an Agentic Data Cloud architecture.
Google Cloud Next '26 highlights the concept of the agentic enterprise, with the new Gemini Enterprise Agent Platform leading the announcements. This comprehensive platform replaces the standalone Vertex AI roadmap and serves as the central hub for building, scaling, governing, and optimizing AI agents. It offers cloud teams lifecycle control rather than just chatbot functionality, integrating model selection, agent building, DevOps, and security into a single system.
Infrastructure updates play a critical role in supporting these production agent deployments across hybrid and multi-cloud environments. Google introduces two new eighth-generation Tensor Processing Unit chips and a specialized AI networking fabric to handle demanding AI workloads. Additionally, the company unveils cross-cloud compute and connectivity updates alongside an Agentic Data Cloud architecture designed to manage data access and operational visibility.
The new platform provides access to more than 200 models through Model Garden, including Google's own Gemini 3.1 Pro and Gemini 3.1 Flash Image, as well as third-party models like Anthropic's Claude family. Google organizes the platform around four core areas: build, scale, govern, and optimize, featuring new tools like Agent Studio and an upgraded Agent Development Kit. These combined hardware and software advancements aim to give organizations the foundational systems needed to run and secure agents at scale.