Broadcom Unveils AgentMinder for Runtime Governance and Identity Control of Enterprise AI Agents

At VMware Explore 2026, Broadcom introduced AgentMinder, a dedicated enterprise control plane and runtime gateway engineered to govern autonomous artificial intelligence agents across private, hybrid, and public cloud environments. Rather than relying strictly on statistical content filtering or human identity credentials, AgentMinder establishes an independent authorization layer positioned between AI agents and enterprise resources. The system assigns certificate-based identities to agents, validates every tool invocation and data request against declared mission parameters and risk profiles in real time, and logs end-to-end execution data. The industry shift from passive language models to autonomous agentic architectures introduces severe governance and security complexities. When agents are empowered to execute database transactions, trigger third-party APIs, and automate business processes, traditional identity and access management (IAM) perimeters break down. Misaligned, compromised, or drifting agents running under broad permissions can cause unintended data exfiltration or unauthorized system modifications at machine speed. AgentMinder tackles this trust gap by decoupling policy enforcement from underlying models, enforcing dynamic least-privilege boundaries at every decision step, and maintaining complete auditability. This release reflects a broader paradigm shift in enterprise AI governance toward "governance by design" and infrastructure-level controls. As global frameworks such as the EU AI Act and NIST AI RMF mandate continuous risk management, verifiable human oversight, and post-deployment monitoring, organizations require programmatically enforceable guardrails. Broadcom's adoption of standard protocols—leveraging the AuthZEN standard for policy interoperability and OpenTelemetry for transparent telemetry streams—signals that AI governance is rapidly embedding directly into core DevOps, DevSecOps, and cloud infrastructure pipelines. For enterprise architects, DevOps practitioners, and platform engineers, operationalizing autonomous agents requires establishing deterministic boundaries before connecting agents to operational systems. Teams should move away from sharing static user credentials with AI workers, transitioning instead to granular machine identities and pre-approved tool manifests. While deploying intermediary policy gateways introduces minor latency overhead during tool resolution, it provides the essential telemetry, attribution, and kill-switch capabilities required to safely run autonomous digital agents at enterprise scale.