Add support for Karpenter as a node provisioning backend in cluster-api-provider-aws

[moko-poi]

/kind feature

Describe the solution you'd like
Currently, cluster-api-provider-aws supports EKS Managed Node Groups (via AWSManagedMachinePool) as a node provisioning option.
However, it does not provide a way to integrate with Karpenter, which is increasingly recommended by AWS and the Kubernetes community for dynamic, workload-driven node provisioning.

I propose adding support for Karpenter as a node provisioning backend, similar to how Managed Node Groups are supported.
For example, a new template type (e.g. AWSKarpenterMachinePool or KarpenterProvisionerTemplate) could allow CAPI-managed clusters to declaratively define Karpenter resources such as Provisioner and AWSNodeTemplate.

This would allow CAPI users to:

• Use declarative infrastructure (via clusterctl, ClusterClass, etc.) with Karpenter-enabled clusters.
• Mix and match static node pools (MGN) and dynamic, reactive node provisioning (Karpenter) under a unified management plane.
• Retain consistency with CAPI’s MachineProvider abstractions for reconciliation, status, upgrades, etc.

Because AWS itself, in its EKS Best Practices documentation, recommends Karpenter as a preferred method for node lifecycle automation and cost optimization, adding first-class support for Karpenter in CAPA would help align CAPI users with AWS’s recommended patterns. :contentReference[oaicite:0]{index=0}

Anything else you would like to add:

• There is an existing experimental project, karpenter-provider-cluster-api, exploring similar ideas, but it is still in PoC stage.
• I am happy to contribute or help with a design proposal or prototype implementation if maintainers are interested.
• Including this support could lower the barrier for CAPI users to adopt more modern, autoscaling node strategies on EKS.

Environment:

• Cluster-api-provider-aws version: (e.g. v2.x.x)
• Kubernetes version: (use kubectl version)
• OS (e.g. from /etc/os-release): (e.g. Amazon Linux 2, Ubuntu 22.04, etc.)

[k8s-ci-robot]

This issue is currently awaiting triage.

If CAPA/CAPI contributors determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.