Google Cloud, Wiz Unveil Agentic Defense at Next ’26

At the Google Cloud Next ’26 conference, Google announced the integration of its multi-billion dollar acquisition, Wiz, into its security stack to build what it calls “Agentic Defense.”

As AI transitions from simple chatbots to autonomous agents that can browse the web and access company data, traditional security isn’t enough. Google’s new Agentic Defense platform combines its own massive threat intelligence with Wiz’s cloud security graph.

The goal is to stop reasoning drift, a phenomenon where an AI agent might deviate from its intended instructions due to malicious prompts or unforeseen data inputs, before it can cause damage to a business.

One of the biggest announcements is the expansion of Wiz’s visibility across the entire development process. A new feature called AI-BOM (Artificial Intelligence Bill of Materials) acts as a detailed ingredient list for AI models. It helps developers track exactly what data and code went into a model, making it easier to spot vulnerabilities or shadow AI i.e. unauthorized AI tools being used within a company.

Wiz is also launching AI Security Hooks. These allow security scanning to happen directly inside coding environments. If an AI agent generates code that contains a security flaw or a hard-coded password, the hook catches it instantly, before the code is even committed to the project

While Google Cloud is the host, the reality for most businesses is a multi-cloud world. The integration with Wiz allows for a single pane of glass view. Security teams can now monitor AI agents and cloud apps across Google Cloud, AWS, Azure, and Oracle Cloud from one dashboard.

This includes new support for agent studios like AWS Agentcore and Salesforce Agentforce. By mapping out every connection between an AI agent and the data it touches, Wiz can identify potential lateral movement paths that a hacker might try to exploit.