Google's enterprise AI push

Google used its Cloud Next event on April 22 to turn Gemini from a model into a full enterprise stack, pairing new chips with new agent-building software. (blog.google) At the center of the launch were two eighth-generation Tensor Processing Units, or TPUs: TPU 8t for training large models and TPU 8i for running them in production. Google said TPU 8i is tuned for fast agent responses, while TPU 8t is built to train complex models on a single large memory pool. (blog.google) A TPU is Google’s in-house artificial intelligence chip, built to handle the matrix math behind model training and inference more efficiently than general-purpose processors. Google said the new chips are designed for “the agentic era,” where software is expected to plan and execute multi-step tasks instead of only answering prompts. (blog.google) Google also introduced Gemini Enterprise Agent Platform, which folds Vertex AI’s model and agent tools into a system for building, deploying, governing and optimizing enterprise agents. The company said the platform adds orchestration, DevOps, integrations and security controls aimed at production use inside large organizations. (cloud.google.com) Agent Studio is the interface layer for that push: a workspace where companies can build agents with Google models, connect them to cloud services and adjust settings such as structured outputs and safety filters. Google’s documentation says Agent Studio also supports deployment and web-app publishing from the same environment. (docs.cloud.google.com) Google is also pushing Gemini into products customers already buy. In Google Ads, Ads Advisor is now described by Google as a Gemini-based conversational assistant, and outside reporting this week said Google added real-time policy troubleshooting, 24/7 security monitoring and instant certifications. (support.google.com) (storyboard18.com) The company is pairing those product updates with reference customers and large contracts. Merck and Google Cloud said on April 22 that they had signed a multi-year partnership, valued at up to $1 billion, to use Gemini Enterprise across research and development, supply chain, employee experience and other operations. (merck.com) (googlecloudpresscorner.com) That combination of custom chips, cloud tooling and embedded distribution is the shape of Google’s enterprise AI pitch in 2026. Instead of selling a single chatbot, Google is trying to sell the infrastructure, the development layer and the everyday business software that puts agents in front of paying customers. (cloud.google.com 1) (cloud.google.com 2)