Cloud platforms are quietly changing their compute foundation. Between traditional virtual machines and containers, a new substrate is emerging: MicroVMs.

MicroVMs combine strong VM isolation with startup times, density, and lifecycle characteristics closer to containers. But their importance goes beyond making serverless platforms faster.
They are increasingly enabling a new execution model: isolated, stateful, disposable sandboxes that can be created on demand, used to safely execute workloads, suspended, resumed, and eventually discarded.

This becomes particularly important in the age of AI agents, where platforms need to execute untrusted or AI-generated code, maintain execution state, provide filesystem and networking isolation, and dynamically provision thousands of short-lived environments.

In this session, we will explore how MicroVMs are being adopted across Azure services and how they are enabling a new generation of compute experiences from serverless and managed container platforms to developer environments, AI sandboxes, and agentic workloads.