Why Microsoft is betting on temporary identities to stop autonomous agents from going rogue

Microsoft is advocating temporary, tightly scoped identities and policy checks to secure autonomous AI agents.

At KubeCon Europe 2026, the company also highlighted Kubernetes fleet management for cloud-to-edge AI operations.

Azure Arc and related tooling are being positioned to simplify deployment and governance of AI workloads across distributed environments.

Microsoft introduced an inference-focused Kubernetes API layer designed to swap underlying engines across environments, improving portability.

The strategy points to a more secure and more portable enterprise model for agentic AI and edge inference.