Your serverless attendee’s guide to AWS re:Invent 2026

extract

AWS Compute Blog Your serverless attendee’s guide to AWS re:Invent 2026 AWS re:Invent 2026 offers an extensive selection of serverless and application integration content. This year’s lineup includes new AWS Lambda execution models: durable functions, Lambda Managed Instances, and Lambda MicroVMs. It also covers AI agents as a new kind of serverless consumer and maturing patterns for production event-driven applications. Don’t have a ticket yet? Register for re:Invent 2026 and join us in Las Vegas from November 30 to December 4. For detailed descriptions and schedule, visit the AWS re:Invent Session Catalog. Sessions are held across several venues, so allow extra time to travel between properties. We’ve curated this guide to help you navigate the serverless sessions and find the content most relevant to you. Session types - Breakout Sessions are lecture-style presentations covering architecture, best practices, and deep dives into AWS services. - Workshops are 2-hour hands-on sessions where you work through tasks in AWS accounts using AWS services. Laptops are required. - Chalk Talks are highly interactive 60-minute sessions with smaller audiences, focused on technical deep dives with whiteboards for architectural discussions. - Builders’ Sessions are 60-minute small-group sessions led by an AWS expert who guides you through a technical problem using AWS services. - Code Talks are 60-minute live coding sessions where AWS experts show how to build solutions using AWS services. - Lightning Talks are 20-minute sessions focused on a single idea. What’s new in AWS Serverless Start here to get oriented on what has changed in serverless heading into re:Invent 2026. SVS207 | Modernizing your compute stack with Lambda’s new execution models – Breakout Session Choose a Lambda execution model for each workload class, run mixed deployments, and compare monitoring across models. Bring your workload portfolio. SVS305 | What’s new in AWS Serverless – Workshop Build with the newest serverless releases and work through the trade-offs that matter in real systems. SVS339 | What’s new with Amazon API Gateway – Breakout Session See what’s new in Amazon API Gateway, including direct service integrations that replace Lambda glue code. Learn how to design APIs for both developers and AI agents, including Model Context Protocol (MCP) servers. Getting started with serverless Are you new to serverless or taking your first steps? These sessions cover fundamentals and hands-on introductions. SVS202 | Building your first event-driven application – Builders’ Session Build a retail order system on Amazon EventBridge, Lambda, and Amazon DynamoDB where each step runs independently and one failure doesn’t cascade. No event-driven experience required. SVS301 | AWS Serverless developer experience – Workshop Build a realistic serverless solution while covering event-driven principles, observability, open-source tools, and deployment automation, in Python, TypeScript, Java, or .NET. SVS302 | Building serverless applications with Terraform – Workshop Define, deploy, and maintain serverless architectures with Terraform, including reusable modules, state management, and deployments across environments. New Lambda execution models Three Lambda execution models cover different workload shapes: durable functions, Lambda Managed Instances, and Lambda MicroVMs. Lambda durable functions API314 | A closer look at AWS Lambda durable functions – Chalk Talk Learn how checkpointing and replay-based recovery work, see patterns such as sagas, fan-out/fan-in, and human-in-the-loop waits, and get a framework for choosing between durable functions and AWS Step Functions. API316 | Coding an AI agent from scratch with AWS Lambda durable functions – Code Talk Build an AI agent from scratch with durable functions, including a tool-use loop, retries with exponential backoff, and human approval that doesn’t bill for idle time. SVS325 | Distributed transactions with Step Functions and durable functions – Builders’ Session Build a payment processing flow with the saga pattern, using Step Functions and durable functions for compensation when part of a transaction fails. Lambda Managed Instances SVS332 | Lambda Managed Instances: Always-on serverless compute – Chalk Talk Lambda Managed Instances offer Amazon Elastic Compute Cloud (Amazon EC2) economics with multi-concurrency while Lambda manages instance lifecycle, patching, and scaling. Model your break-even point, and bring your Lambda bill. Lambda MicroVMs SVS307 | Running untrusted code in sandboxes with AWS Lambda MicroVMs – Builders’ Session Build sandboxes that run untrusted code, such as AI-generated code and third-party plugins, with resource limits, permission boundaries, and tested failure modes. SVS337 | AWS Lambda MicroVMs: Strong isolation for agents and beyond – Breakout Session Learn how Lambda MicroVMs work: VM-level isolation per agent, tenant, or task, with state that persists across invocations and sandboxes that suspend when idle and resume on demand. Serverless and AI agents A major theme this year is building, orchestrating, and operating AI agents on serverless infrastructure. SVS338 | Serverless developers in the agentic era: What changes, what doesn’t – Breakout Session See what changes when AI agents call your functions as tools, and what stays the same. Topics include orchestration for non-deterministic workflows and cost governance when one reasoning chain calls 200 functions. SVS306 | Building agentic AI architectures with AWS Serverless – Builders’ Session Build a multi-agent system that uses Lambda for tool execution, Amazon Bedrock for reasoning, and Amazon EventBridge for communication between agents. SVS324 | Building agentic apps with serverless: Lambda as your agent’s toolbox – Chalk Talk Design Lambda functions as agent tools, with single-responsibility tools, clear input and output schemas, MCP-compatible interfaces, and multi-tenant isolation. SVS342 | Coding the tool layer for AI agents on serverless – Code Talk Retrofit a serverless API into an agent tool, live. The build covers an MCP server with versioned tools, discovery through the Gateway capability of Amazon Bedrock AgentCore, a Step Functions workflow exposed as one tool, and scoped agent identity. API312 | Building event-driven, multi-agent architectures on serverless – Chalk Talk Map agent coordination patterns on Amazon EventBridge, such as peer-to-peer messaging, supervisor-worker hierarchies, and broadcast, and the failure modes they introduce, including circular delegation. COM325 | The agent orchestration spectrum: agents, Step Functions, or Lambda? – Code Talk Live-code the same AI workload with AgentCore, Step Functions, and durable functions, and compare latency, cost, and debuggability on a production content platform. SVS328 | Guardrails for AI-generated serverless infrastructure – Chalk Talk Coding agents can generate over-permissioned AWS Identity and Access Management (IAM) policies and skip encryption. Learn which guardrails to automate and which need human review. API402 | Building event-driven architectures in the age of AI agents – Breakout Session See how familiar patterns such as queues, events, and retries evolve when your consumers are autonomous agents instead of microservices. Building event-driven architectures Event-driven architectures remain the backbone of serverless applications. These sessions cover patterns from foundational to advanced. API302 | Serverlesspresso: Build an event-driven app from the ground up – Workshop Build a serverless coffee-ordering app with Lambda, Amazon EventBridge, and durable functions, which checkpoint progress and recover from failures without custom state management. API305 | Building enterprise-scale event buses with Amazon EventBridge – Builders’ Session Configure an event bus with routing policies, schema discovery, and access controls so platform teams govern events while application teams ship independently. API311 | How event-driven architectures go wrong and how to fix them – Chalk Talk Dissect costly anti-patterns such as unbounded fan-out, missing dead-letter queues, chatty event schemas, and orchestration gaps that silently drop work, and fix them. API317 | Coding event-driven retry patterns that prevent outages – Code Talk Live-code exponential backoff with jitter, circuit breakers, poison-pill isolation, and replay from archive for Amazon EventBridge and Step Functions. Orchestration: Step Functions and durable functions Both Step Functions and Lambda durable functions handle retries, long-running execution, and failure recovery. These sessions help you choose. API318 | Composing the right orchestration for serverless workflows – Lightning Talk A decision framework for human approvals, long-running AI agent workflows, and multi-service sagas, plus a hybrid pattern that uses durable functions inside Step Functions. ARC329 | The coordination decision: when to orchestrate and when to let go – Breakout Session Decide per flow whether to orchestrate or let services react to events, using a logistics platform that runs both patterns on Step Functions and Amazon EventBridge. Building APIs and serverless web applications Serverless is increasingly used to build full-stack web applications, not only API backends. SVS348 | Serverless compute built for the web – Breakout Session Build full-stack web applications entirely on serverless services, with patterns for HTTP routing, server-side rendering, static assets, and real-time features. SVS327 | Build interactive web apps with AWS Lambda – Chalk Talk Run web frameworks such as Express, Hono, and Next.js on Lambda with runtime adapters, including migration paths for containerized web apps. API304 | API Gateway unleashed – Workshop Create, deploy, and manage APIs with API Gateway, including message transformation, request validation, security, Infrastructure as Code, and WebSocket backends. API306 | Build secure APIs faster with Amazon API Gateway and Kiro CLI – Builders’ Session Implement and compare API Gateway authorization with Amazon Cognito user pools, Amazon Verified Permissions, Lambda custom authorizers, and IAM authorizers. Then go from spec to deployed API with Kiro CLI, OpenAPI, and MCP servers. API309 | Amazon API Gateway: Migration strategies and agent-ready APIs – Chalk Talk Learn migration strategies from legacy API platforms, then design APIs that AI agents can discover, authenticate against, and call, including MCP-compatible tool definitions. Serverless developer experience AI-assisted development is changing how serverless applications are built, tested, and deployed. SVS303 | AI-driven serverless development with Kiro – Workshop Use Kiro to build a serverless application with authentication, tests, observability, and an optional continuous integration and continuous delivery (CI/CD) pipeline. SVS343 | A day in the life of a serverless builder – Code Talk Follow one serverless application from first line of code to production, live: an AI coding agent generates the scaffolding, and you debug a failing request with OpenTelemetry. SVS329 | Testing serverless and event-driven applications – Chalk Talk Whiteboard testing strategies for serverless and event-driven apps, from unit and contract tests to chaos techniques such as duplicate and out-of-order delivery. SVS322 | Build self-healing serverless applications with AWS DevOps Agent – Builders’ Session Inject failures such as Lambda throttling, timeout cascades, and dead-letter queue buildup, then configure AWS DevOps Agent to find root causes and apply fixes with approval gates. Performance, cost, and operations at scale Tune your Lambda functions, optimize costs, and learn what changes when you operate at enterprise scale. SVS341 | Lambda performance tuning and cost optimization – Chalk Talk Go beyond memory tuning with SnapStart and provisioned concurrency, event source mapping tuning, and guidance on when Lambda Managed Instances beat per-invocation pricing. SVS335 | Operating serverless at scale: What changes at 1,000+ functions – Chalk Talk Learn what changes between 10 and 1,000 functions: safe deployments, account structure and concurrency management, observability without cardinality overload, and platform patterns. ARC325 | Serverless at 1M RPS: Lambda, DynamoDB, and SQS scaling lessons – Chalk Talk Push Lambda, DynamoDB, and Amazon Simple Queue Service (Amazon SQS) to one million requests per second and see what breaks. The session shares real numbers on cold start budgets, retry amplification, and backpressure. SVS334 | Serverless observability: OpenTelemetry and vended logs – Chalk Talk Set up OpenTelemetry and API Gateway vended logs for tracing across Step Functions, Amazon EventBridge, and SQS, plus observability for agentic workloads such as token use. SVS320 | Same code, any compute: Portable architecture on AWS – Builders’ Session Apply Hexagonal Architecture and 12-factor patterns so compute is a deployment decision, not an architectural one, using the AWS Lambda Web Adapter and runtime-neutral handlers. Serverless data processing Process data at scale without provisioning clusters or managing batch infrastructure. SVS309 | Building serverless distributed data processing workloads – Builders’ Session Build a pipeline with Step Functions Distributed Map, Lambda, and Amazon Simple Storage Service (Amazon S3) that processes thousands of objects in parallel and handles partial failures. SVS321 | Serverless streaming data processing with AWS Lambda and Apache Kafka – Builders’ Session Build a real-time location tracking app that ingests events from Amazon Managed Streaming for Apache Kafka (Amazon MSK), processes them with Lambda, and updates a dashboard through AWS AppSync. Community sessions Sessions led by AWS community members and customers sharing production experience with the preceding execution models and patterns. COM303 | Lambda for every scale and workload shape – Breakout Session Honeycomb shares how it runs its query backend across all three Lambda execution models. Honeycomb reports that Managed Instances absorb steady-state traffic at roughly 7.5x lower cost than Functions, while MicroVMs take spikes and Functions catch bursts. COM326 | Observable agents with Lambda durable functions and OpenTelemetry – Code Talk Build a multi-agent purchasing coordinator on durable functions and instrument it with OpenTelemetry for one coherent trace across all invocations, plus Amazon CloudWatch insights for fleet-wide queries. COM304 | Hidden trade-offs in distributed architectures – Breakout Session Investigate unexpected system behavior in a real-world case study built on Lambda, DynamoDB, DynamoDB Streams, and Step Functions, and learn the trade-offs behind it. Additional resources Many sessions run more than once, so if a session is full, check the catalog for another time. If you cannot join in person, we upload breakout sessions to the AWS Events YouTube channel after the event. For more serverless learning resources, visit Serverless Land. We look forward to seeing you at re:Invent 2026!