A managed node group is part of an AWS ASG (AutoScalingGroup) and means that the nodes, if they fail health checks, will be rescheduled.
They also get some extra annotations that mean if you deploy Karpenter or the AWS Cluster Autoscaler, the node group will be scaled up and down as required.