A year ago, choosing a container hosting option on Azure was already confusing. Azure Container Instances, Azure Container Apps, Azure Kubernetes Service, App Service, Azure Functions. Five different answers to the same question. Ask three developers which one to use and you get five different opinions. Microsoft assessed the situation and concluded that what was clearly missing was more options.

Azure Container Apps now comes in three distinct tiers. There is serverless GPU support with true scale-to-zero billing. Dynamic Sessions let you run AI-generated code in Hyper-V isolated sandboxes. And Azure Kubernetes Fleet Manager has matured into a standalone service for orchestrating entire fleets of clusters. The line between "I just want to run a container" and "I need a full orchestration platform" has never been harder to see. Congratulations to everyone involved.

This session maps the full landscape. I evaluate hosting options side by side across dimensions that actually matter: complexity, scalability, cost, cold start behavior, and how quickly you can get something running without filing a support ticket. Whether you are hosting microservices, a background job, AI workloads, or a full multi-tenant platform, you will leave with a clear decision framework for your specific situation.

No marketing, no hand-waving. Just an honest look at the full landscape and the trade-offs involved, so that the next time someone asks which container service to use, you will have a concrete answer. Not five opinions. One answer. Perhaps Microsoft already anounced another hosting option by the time you get home. At least you'll know where it fits in the landscape