AI infrastructure from cloud to edge | INV643

Mark Russinovich and Pavan Davuluri present how Azure and Windows support production AI across cloud, datacenter, edge, and device environments, with demonstrations of factory inference and application modernization.

Azure infrastructure for production AI

The session examines infrastructure for scaling AI workloads and agents, positioning Azure as the foundation for mission-critical AI systems. It covers infrastructure innovations for scale and security, followed by new AI primitives:

The enterprise workload discussion includes AKS AI Manager and Foundry integration. A factory demonstration brings together AKS, Foundry, and edge inference to show an AI scenario aimed at reducing downtime.

AI-driven application modernization

The session introduces Azure Catalyst for AI-driven application modernization. A workflow demonstration covers modernization with migration automation.

Security and deployment across environments

The security discussion covers confidential computing, sovereign cloud, and AI deployment across environments. The Windows portion focuses on hybrid intelligence and enterprise security, followed by agent security, local AI on Windows, Copilot innovations, and Windows Server 2028.

Together, these topics address running intelligence where it provides value: in Azure, in datacenters, at the edge, and across more than a billion Windows devices.

Session chapters

Session details and resource

This is Microsoft Ignite 2026 session INV643, a foundational-level Innovation Session, presented in English (US) and categorized under Windows.