Learn how enterprises can protect sensitive AI data and code through Confidential AI in this 21-minute Microsoft Ignite conference talk. Explore how Ubuntu confidential VMs powered by AMD 4th Gen EPYC processors and NVIDIA H100 GPUs on Microsoft Azure provide hardware-rooted execution environments for both CPU and GPU workloads. Discover key security concepts including memory confidentiality, integrity, and remote attestation that help organizations overcome data privacy and intellectual property concerns in cloud-based machine learning. Presented by speakers Jehudi Castro-Sierra, Antoine Delignat-Lavaud, and Ijlal Loutfi, gain practical insights into implementing secure AI workloads that protect valuable private data.