AWS introduces Lambda Managed Instances (LMI), a new capability allowing Lambda functions to run on dedicated Amazon EC2 instances like Graviton4. This hybrid approach eliminates managing instance lifecycles, patching, and scaling policies, while offering access to specialized compute options and EC2 pricing advantages like Savings Plans.

LMI is designed for predictable, high-volume workloads and performance-critical applications needing specific CPU or network profiles. It provides automatic provisioning and dynamic scaling, handling traffic doubling within five minutes. This model suits regulatory needs requiring granular VPC control, blending serverless convenience with the power of dedicated infrastructure.

A sample application demonstrates the architecture using Java 25, AWS SAM, Amazon API Gateway, and Amazon DynamoDB. The code, available on GitHub, shows Lambda handlers for creating and retrieving products. It uses the AWS SDK for Java to interact with DynamoDB, with the sample secured by an API key for endpoint access.