"We’re bringing zero trust verification to every agent and at every orchestration step," Thomas Kurian, Google Cloud’s CEO, said at Google Cloud Next 26 — a concise pledge that frames Google’s newest push to govern autonomous software.
Unique cryptographic IDs, Agent Registry and traceable policies
Google’s Gemini Enterprise Agent Platform introduces a persistent identity for AI agents: each agent is assigned a unique cryptographic ID that is recorded and referred to for every action the agent takes. According to Google, those agent IDs are designed to be mapped back to "defined authorization policies that are traceable and auditable." The platform includes an Agent Registry, a central library that indexes every internal agent, tool and skill, giving organizations a single place to catalog the autonomous actors running against their systems.
Agent Gateway, MCP and A2A support, and Model Armor protections
The platform also offers an Agent Gateway — a single dashboard intended to manage a fleet of AI agents and enforce policies for agent-to-agent and agent-to-tool connections. Google said the Gateway supports several agentic AI protocols such as the model context protocol (MCP) and Agent2Agent (A2A). In an April 22 statement, Google Cloud described the Gateway as providing "secure, unified connectivity between agents and tools across any environment, while enforcing consistent security policies and Model Armor protections to safeguard against prompt injection and data leakage." Model Armor is presented as Google Cloud’s guardrail layer against adversarial attacks, prompt injection, sensitive data leaks and harmful content.
Agent Anomaly Detection, Agent Threat Detection, and the Agent Security dashboard
At Cloud Next 26, Google unveiled Agent Anomaly Detection, a feature that uses statistical models and "a large language model (LLM) as-a-judge framework" to identify unusual behavior in real time — flagging suspicious reasoning patterns and other deviations from expected operation. That capability is intended to work alongside existing Agent Threat Detection, which the company says monitors malicious activities such as reverse shells and connections to known bad IP addresses.
These detection tools feed into a new Agent Security dashboard, powered by Google Cloud’s Security Command Center (SCC). Google Cloud said the dashboard unifies threat detection and risk analysis within Google Cloud Platform environments, helping security teams map relationships between AI agents and models, automate asset discovery and scan for vulnerabilities in operating systems and language packages. The dashboard builds on Gemini Enterprise features such as Agent Compliance and Agent Policy, which already provide policy enforcement.
New cybersecurity agents, dark web intelligence and AI infrastructure
Google also announced purpose-built agents for security work. Two agents — a Threat Hunting agent and a Detection Engineering agent — are available in preview; the former is intended to help security teams hunt for novel attack patterns and stealthy adversary behaviors, while the latter aims to identify coverage gaps and create new detections for threat scenarios. A Third-Party Context agent is "coming soon to preview" and is intended to enrich workflows with contextual data from third-party content. When available, Google says these agents will be integrated into Google Security Operations, the firm’s security analytics, threat detection and incident response platform.
Google highlighted earlier gains from its agent work: the Triage and Investigation agent, introduced in April 2025, "processed over five million alerts in the last year," a workflow Google said reduced a "typical 30-minute manual analysis to 60 seconds." Separately, a new dark web intelligence feature in Google Threat Intelligence is available in preview; Google reported internal tests showing the feature can analyze millions of daily external events with 98% accuracy to elevate threats that matter.
On the hardware and partner side, Google released two AI-focused processing chips — the Tensor Processing Unit 8t (TPU 8t) for AI training and the Tensor Processing Unit 8i for AI inference — and committed $750m to an agentic AI partner fund aimed at global consulting firms, systems integrators, software partners and channel partners for agent prototyping, deployment and upskilling.
What this means for security teams, developers, and procurement partners
- Security teams: They gain centralized visibility (Agent Registry and Agent Security dashboard), automated anomaly and threat detection capabilities, and agent-level identities that map to auditable authorization policies — tools Google presents as designed to reduce manual analysis and increase traceability.
- Developers and DevSecOps: Google’s integration of Wiz’s AI-Application Protection Platform (AI-APP) — expanded to embed real-time vulnerability scanning, AI-generated code security, a dynamic AI bill-of-materials (AIBOM) and automated remediation into platforms like AI development solution Lovable, IDEs and version control systems — signals a push to move security earlier in agent development workflows.
- Procurement partners and integrators: The $750m agentic AI partner fund targets global consulting firms, systems integrators and channel partners to support AI value identification, agent prototyping and upskilling, making partner-led projects a clear channel for enterprise adoption of Gemini Enterprise agent tools.
Google framed the Gemini Enterprise Agent Platform as an attempt to make a new class of dynamic digital entities — autonomous, goal-oriented AI agents — as manageable as traditional non-human identities by assigning cryptographic IDs, enforcing zero-trust verification and layering detection and audit tools across orchestration steps. Announced at Cloud Next 26 (April 22–24), these capabilities are rolling out as previews and integrations; they underline Google Cloud’s bet that governance, traceability and embedded security will define enterprise adoption of agentic AI.
https://www.infosecurity-magazine.com/news/google-ai-agent-identities-gemini/
