Cast AI for Karpenter is now generally available, adding an operational layer on top of Karpenter for EKS clusters. It addresses gaps that Karpenter alone doesn't solve: workload-level cost visibility, continuous cluster rebalancing, Spot interruption prediction up to 60 minutes in advance, and automatic alignment of pod resource requests with real usage via a Workload Autoscaler. The integration is non-invasive — it connects in read-and-recommend mode by default, preserves existing Karpenter configurations, and supports GitOps workflows via Helm. Automation is opt-in per workload and cluster.
Table of contents
Faster provisioning creates a new operating challengeWith Cast AI, Karperter now has an operational layerStart safely, then automate at your paceWhen Cast AI for Karpenter makes senseGet started with Cast AI for Karpenter nowSort: