Confidential containers on Azure Container Instances

They are Azure Container Instances designed to protect container workloads from the host operating system by encrypting virtual-machine memory. The technology uses AMD SEV-SNP to create a confidential-computing boundary around the workload.

AMD SEV-SNP helps protect data and code while they are in use inside encrypted VM memory. This is useful when a workload processes sensitive information and the operator wants stronger isolation from the underlying host.

They use the same general container-instance model, but add confidential-computing protection through a specialized hardware-backed environment. Application compatibility, supported regions, image requirements, and networking options must be checked before migration.

They are useful for sensitive analytics, regulated data processing, confidential machine-learning workloads, and multi-party computation. Encryption at rest and in transit still matters because memory encryption does not remove the need for normal identity, network, and storage controls.