From models to infrastructure: Confidential AI at Hyperscale on Azure | BRK443
Sai Deep Tetali and Vikas Bhatia present an intermediate-level session on confidential AI for organizations deploying large AI workloads in Azure.
Confidential AI on Azure’s VR72 platform
The session examines how Azure Confidential VMs on the next-generation VR72 platform support frontier-scale AI workloads. Its scope extends from protecting AI models and their inputs to protecting the infrastructure that runs them.
Key areas include:
- Hardware-backed protection: Protection for models, prompts, data, and infrastructure.
- Performance: Running protected AI workloads without compromising performance.
- Scale: Meeting the demands of frontier-scale AI deployments.
The central focus is bringing these security, performance, and scale requirements together rather than treating protection as a separate concern.
Session details and resources
- Event: Microsoft Ignite 2026
- Session: BRK443 — Cloud & AI
- Format: Breakout
- Level: Intermediate (200)
For further learning, see the Cloud & AI next-step resources.