**What are Serverless Deployments in Microsoft Foundry?**
Serverless deployments (also called serverless API or Models-as-a-Service) let you use AI models in Microsoft Foundry through an API endpoint without provisioning or managing any compute infrastructure. Microsoft hosts and scales the model for you, and you pay per usage (typically per token) rather than for dedicated VMs. It’s ideal for quickly integrating models into apps with low setup effort and variable workloads.