Amazon Web Services Launches Two New Amazon Lambda Features to Better Address Complex Application Scenarios

At the 2025 AWS re:Invent global conference, AWS announced two new Amazon Lambda features—Amazon Lambda Durable Functions and Amazon Lambda Managed Instances. These capabilities address long-running workflows requiring progress preservation across interruptions and high-performance scenarios demanding precise control over underlying infrastructure, enabling developers to handle more complex applications and business needs while maintaining the simplicity of serverless architecture.

Amazon Lambda Durable Functions empower reliable multi-step applications and AI workflows

Over 1.8 million customers use Amazon Lambda each month to process more than 15 trillion requests. Developers rely on Amazon Lambda to quickly build functions that run code, and they also seek to build multi-step applications that can execute reliably over extended periods—such as payment processing, customer onboarding flows, or AI workflow orchestration. Amazon Lambda Durable Functions allow customers to create Lambda functions that preserve execution state even when interrupted, supporting pauses of up to one year.

Amazon Lambda tracks function progress and ensures reliable execution through built-in error handling and automatic recovery mechanisms, even during pauses or disruptions. Now, developers can use Amazon Lambda to build robust, long-running, multi-step applications and AI workflows, unlocking entirely new use cases.

Amazon Lambda Managed Instances extend the benefits of serverless architecture

Customers widely adopt Amazon Lambda for building serverless applications thanks to its simple development experience, automatic scaling, and intuitive operational model. However, for use cases such as data analytics pipelines, financial risk modeling, and multiplayer gaming, customers also require consistent large-scale compute performance and fine-grained infrastructure control. Amazon Lambda Managed Instances bridge the gap between the simplicity of serverless and the need for infrastructure control by allowing customers to run Lambda functions on their chosen Amazon EC2 instances.

Customers can now select compute resources tailored to their workloads and immediately benefit from the latest Amazon EC2 hardware and pricing models. Whether processing video, running complex algorithms, or handling other demanding tasks, developers can set target utilization levels for their compute resources. AWS automatically adjusts the size of the instance fleet based on demand, while developers continue building with the familiar Amazon Lambda programming model. With Amazon Lambda Managed Instances, developers gain full control over performance characteristics, delivering optimal experiences for end users.