Google bets on agent security
- Google announced a broad security offensive built around AI agents and the security firm Wiz. - The company paired a planned $32 billion acquisition of Wiz with a $750 million agentic AI partner fund. - It unveiled an "Agentic Defense" capability to detect shadow AI and automate threat response, treating agents as new attack surfaces. ( )
Google is tying its cloud security push to AI agents, folding Wiz into a broader plan to spot and contain machine-speed attacks. (cloud.google.com) At Google Cloud Next on April 22, 2026, the company said its new “Agentic Defense” offering combines Google Threat Intelligence and Google Security Operations with Wiz’s cloud and AI security platform. (cloud.google.com) Google had already closed its $32 billion acquisition of Wiz on March 11, 2026, after first announcing the deal in March 2025, and said Wiz will keep its brand and continue supporting Amazon Web Services, Microsoft Azure, Oracle Cloud, and Google Cloud. (blog.google) The basic problem is that AI agents are software workers with permissions, data access, and the ability to trigger actions across systems. Google said that shift creates “a new set of threats” aimed at both AI models and the business data used to give those agents context. (cloud.google.com) Google is also spending to make more of those agents. On April 22, it announced a $750 million partner fund for agent development and deployment, covering consulting firms, software vendors, and channel partners in its ecosystem. (cloud.google.com) That fund sits alongside Google’s Gemini Enterprise Agent Platform, which includes an Agent Gallery for vetted third-party agents and new ways for customers to deploy partner-built tools inside Gemini Enterprise. (cloud.google.com, cloud.google.com) Google’s pitch is that companies now need to secure both the agents they deploy and the attacks those agents may face or amplify. The company said the combined Google-Wiz platform is meant to cover code, cloud, runtime, and AI systems across hybrid and multicloud environments. (blog.google, cloud.google.com) Outside Google, the argument is that defenders are trying to match attackers’ speed with more automation. ZDNET described the new portfolio as AI that can “hunt, detect, and fix” threats faster than human teams working alone. (zdnet.com) Google has been building toward this framing for months. At RSA Conference in March, it said “agentic AI defense” would automate more detection, triage, and response as security teams confront faster, AI-assisted intrusions. (cloud.google.com) The bet is that the same companies buying AI agents to do work will also need tools to govern them, watch for “shadow” deployments, and lock them down across multiple clouds. Google is now trying to sell both sides of that equation. (cloud.google.com, cloud.google.com)
