On June 11, 2026, JumpCloud said it is launching Agentic IAM on Google Cloud, extending its identity and device-management platform to govern access for human users, non-human identities, and autonomous AI agents. The service is hosted on Google Cloud, optimized for Gemini Enterprise, and positioned as a centralized control plane for discovering, registering, authenticating, and auditing agent access across enterprise systems.
The timing matters because enterprise AI governance is moving closer to runtime. As more agents start calling tools, touching SaaS systems, and routing work across MCP and agent-to-agent flows, older IAM models built around human users and static service accounts are starting to look incomplete.
What JumpCloud launched on June 11
This was not framed as a generic AI dashboard. JumpCloud is pitching Agentic IAM as an identity layer for agents that operate at machine speed and can create a much larger blast radius than a typical bot or integration account.
- Real-time discovery and registration: JumpCloud says the platform is designed to discover and register agent identities as they appear across the enterprise.
- Gemini Enterprise optimization: The launch is specifically tied to Google Cloud and Gemini Enterprise, giving JumpCloud a clearer route into organizations standardizing on Google’s agent stack.
- AI Device Trust: JumpCloud says device trust checks can verify the human operator behind AI requests, including biometric and environment-based checks for Gemini Enterprise sessions.
- Unified AI gateway controls: The company says MCP and agent-to-agent traffic can be governed through OpenID Connect-based flows and short-lived token controls.
JumpCloud also said the rollout will continue through 2026, including Managed AI Connectors and broader AI Device Trust capabilities for Gemini Enterprise sessions.
Why the Google Cloud piece matters more than simple hosting
The deeper signal is that Google Cloud is increasingly treating agent governance as infrastructure, not a bolt-on feature. Its IAM portfolio now explicitly covers users and agents together, while its Gemini Enterprise Agent Platform includes Agent Identity, Agent Registry, and Agent Gateway as core control surfaces.
That matters for JumpCloud because enterprise buyers increasingly want one place to answer a new set of operational questions: which agents exist, which tools they can call, which MCP servers are approved, what happens when an agent acts on behalf of a user, and where prompt-injection or sensitive-data risks are stopped before they spread.
Google’s own Agent Gateway documentation points in the same direction. Access can be tied to agent identity, enforced through Identity-Aware Proxy, and extended with Model Armor for runtime protections such as prompt-injection defense and sensitive-data controls. In other words, JumpCloud is aligning itself with a cloud platform that is already building policy, registry, gateway, and identity primitives for AI agents instead of forcing agent governance through older user-centric controls.
Business impact for enterprise AI and security teams
The most important takeaway is that AI identity is becoming its own operating category. JumpCloud cited research saying 66% of organizations now give AI agents equal or greater system access than human users, while only 37% have fully integrated those agents into formal IAM policies. Whether those figures end up matching every enterprise environment or not, the direction is clear: rollout speed is outpacing control maturity.
That creates three immediate pressures for enterprise teams:
- Inventory pressure: security and IT teams need to know which agents already exist, including shadow deployments and unsanctioned connectors.
- Authorization pressure: least-privilege access has to work for fast tool-calling agents, not only for employees signing into applications.
- Runtime pressure: policy enforcement has to stay active after deployment, when agents are making live requests and touching sensitive data.
For CIOs, CISOs, and AI platform leaders, that shifts the buying conversation. The harder question is no longer only which model to use. It is which control plane can prove who an agent is, what it can reach, what it did, and which policies applied when it acted.
What to watch next
The next test is whether JumpCloud can turn this into a durable operating layer rather than a positioning story. Enterprises will want to see how deep the Gemini Enterprise integration becomes, how much policy can be enforced across real MCP and agent-to-agent traffic, and how well Agentic IAM governs third-party agent stacks without creating fresh lock-in.
The broader implication is bigger than one launch. Enterprise AI is entering a phase where identity, runtime authorization, and traffic governance matter nearly as much as model capability. For teams building AI agents, automation systems, and multi-agent workflows, the vendors that win may be the ones that can make agent behavior legible, governable, and auditable in production.