Lambda Managed Instances runs your functions on EC2 instances in your account, many requests per environment, at EC2 prices plus a 15% fee. It removes cold starts and opens Savings Plans to Lambda, and it gives up scale-to-zero. Where it fits between standard Lambda and Fargate.
Standard Lambda bills memory multiplied by wall-clock time, one request per execution environment. For bursty traffic that is the best deal on AWS. For a function serving steady traffic all day, it means paying for every millisecond spent waiting on a database, in every environment, at a rate no Savings Plan touches deeply.
Lambda Managed Instances, launched at re:Invent 2025, changes the model. You create a capacity provider (a VPC, optionally instance requirements and scaling settings), attach functions to it, and Lambda runs them on EC2 instances in your account that it provisions, patches and replaces. One execution environment handles many invocations at once.
Book a free 30-minute call. You'll leave knowing what we'd do first and roughly what it costs, whether or not you hire us.
Your code must be safe under concurrency. Globals, module-level caches and connection objects are now shared between simultaneous requests. A function written for one-request-at-a-time can have bugs that only appear here.
There is a floor. A function version runs on three instances by default even at zero traffic. That is fine for a service that is always busy and wasteful for one that is not.
Scaling is slower than standard Lambda. AWS describes it as handling traffic doubling within five minutes by default. A sudden 20x spike is still standard Lambda's job.
Isolation is different. Standard Lambda isolates each environment in a Firecracker microVM on a shared fleet. Managed Instances run containers on Nitro instances in your account, and the capacity provider is the security boundary. Put functions with different trust levels on different capacity providers.
Standard Lambda: spiky or low traffic, event processing, anything that benefits from scale to zero. Still the default.
Lambda Managed Instances: functions that are busy all day, I/O-heavy APIs where multi-concurrency packs many requests on one environment, and teams with EC2 Savings Plans to spend. You keep Lambda's event sources and deployment model.
Fargate: long-lived servers, WebSockets, anything that is already a container and does not need Lambda's event model. Also the better choice when the team wants one deployment unit for everything.
Once a month: what changed in tech that month and what it means for you, from cloud and AI to security and the tools you build with. Every item links to its source. No sales pitch, unsubscribe in one click.
Book a free 30-minute call about it. Bring your setup and your bill; you'll leave knowing what we'd do first and roughly what it costs, whether or not you hire us.
Or keep reading
Lambda Web Functions vs Event Functions: Where the New Model Saves Money, and Where It Doesn't
Web Functions bill active CPU and held memory; Event Functions bill wall-clock time multiplied by memory. For AWS's own three example workloads, classic Lambda at 512 MB is cheaper. Once requests spend seconds waiting (LLM streaming, slow APIs, long polls) Web Functions win by 40 times or more: a 10-second stream costs about $3 per million requests instead of $134. The arithmetic, the crossover and a checklist.
Read
AWS Lambda Is Five Things Now: Event Functions, Web Functions, Durable Functions, MicroVMs and Managed Instances
In twelve months Lambda grew from one compute model to five. Event Functions (the classic model, renamed), Web Functions for HTTP servers billed on active CPU, Durable Functions for long workflows, MicroVMs for isolated sandboxes, and Managed Instances on EC2. What each is for and how each bills.
Read
As of October 25, 2026. Provided as is, for general information, and accurate to the best of our knowledge on the date shown. Not legal, compliance, tax, financial or other professional advice. Prices, limits and vendor services change; check current vendor documentation before relying on anything here. Terms of use.