AWS adds x86 to Lambda MicroVMs: Arm stays default, four Regions excluded

What the new x86 option means

On November 11, 2026, AWS announced x86 CPU support for Lambda MicroVMs. A MicroVM is a virtual-machine-style execution environment that can start on demand and stay isolated from other environments. AWS lists interactive development environments, CI/CD jobs, AI-generated code sandboxes, and security scanning among its use cases. For teams with existing x86 Linux tools or native dependencies, images can now be created for x86; AWS says existing x86 applications, tools, and software dependencies can run without recompiling or porting.

For a build that already depends on x86 binaries or native libraries, the new option reduces the need to adapt dependencies for an architecture migration. Validate the operating system, software licenses, and dependency combination in the image; the announcement does not provide a performance comparison between architectures.

Choose x86 or Arm for the workload

When creating a MicroVM image from a Dockerfile, you can choose x86 or Arm (AWS Graviton). Arm remains the default. If an existing build depends on x86 binaries, native libraries, or developer tools, assess whether an x86 image lets you reuse them as-is. If the workload already targets Arm, you can continue choosing Arm. The added option does not mean every software package, container image, or external tool will work automatically.

Record the architecture as part of the image build decision and keep it consistent across the CI/CD image and the jobs that use it. To compare architectures, use the same authorized workload and record resource settings and dependency versions so that configuration differences are not mistaken for architecture differences.

Check Region support, including four exclusions

AWS says x86 support is available in all AWS Regions where the service is available, except Asia Pacific (Hong Kong), Asia Pacific (Malaysia), Asia Pacific (Thailand), and Europe (Spain). Hong Kong is therefore outside the announced x86 support scope. This list describes AWS service Regions; it does not determine a user's nationality or physical location, and it does not mean that accounts, capacity, or access requirements are automatically satisfied elsewhere.

Before planning a deployment, verify Lambda MicroVM availability in the target Region and your own AWS account. If the workload needs another supported Region, separately assess network paths, latency, data residency, and organizational policy. The architecture announcement alone cannot establish that those conditions are met.

Create an image from a Dockerfile and plan validation

Prepare the application code and a Dockerfile, package them into a ZIP archive, upload it to Amazon S3, then create a MicroVM image with the required architecture. AWS lists the Lambda console, CloudFormation, AWS CDK, and development tools using the Agent Toolkit for AWS as access routes. Use the architecture field documented by the integration or image-creation interface you choose.

The image captures a snapshot of the initialized environment, and isolated MicroVMs can launch from that snapshot. The guide also says an idle MicroVM can be suspended while preserving memory and disk state, then resumed later. Configure ingress and egress networking for the workload separately. x86 support does not provide a fixed exit IP or guarantee access to a target website or acceptance by an external platform.

Include cost and Region checks before rollout

The AWS Lambda pricing page lists MicroVM charges for compute, snapshot reads and writes and storage, plus standard data transfer. Baseline resources are charged while a MicroVM is running; resources used above the baseline are charged for their active use. The page explicitly allows suspension for up to 8 hours without compute charges while preserving memory and disk state. Snapshot reads, writes, and storage still carry charges, so suspension does not make the entire bill zero.

Snapshots are read on start or resume and written on suspend. MicroVM image storage also has a 1-week minimum retention period. Include these items in the budget even for a small validation run, and check standard data-transfer charges for the Region in use.

Before rollout, confirm whether the workload needs x86 or Arm, verify that the target Region offers MicroVMs, and check the image architecture. Then use a small authorized job to inspect dependencies, ingress and egress, suspend and resume behavior, and estimated costs. Keep the configuration and results in the deployment record to help decide whether to expand usage.