Azure Databricks customers can now specify AMD-based confidential VMs for their Azure Databricks cluster driver node and cluster worker nodes. Azure AMD-based confidential VMs provide a strong, hardware-enforced boundary that hardens the protection of the guest OS against host operator access and other Azure tenants. These VMs are designed to help ensure that data in use, in memory, is protected from unauthorized users using encryption keys generated by the underlying chipset and inaccessible to Azure operators. This will help customers analyze more sensitive data on Azure Databricks.
To learn more, read the announcement blog and the documentation.