diff --git a/api/v1beta1/common_types.go b/api/v1beta1/common_types.go
index d489a2d66896..7137cef0e983 100644
--- a/api/v1beta1/common_types.go
+++ b/api/v1beta1/common_types.go
@@ -22,6 +22,9 @@ import (
 )
 
 const (
+	// TakeOverCluster is the label used to mark the nodes that run on takeover-cluster instances.
+	TakeOverCluster = "cluster.x-k8s.io/takeover-cluster"
+	
 	// ClusterLabelName is the label set on machines linked to a cluster and
 	// external objects(bootstrap and infrastructure providers).
 	ClusterLabelName = "cluster.x-k8s.io/cluster-name"
diff --git a/bootstrap/kubeadm/internal/controllers/kubeadmconfig_controller.go b/bootstrap/kubeadm/internal/controllers/kubeadmconfig_controller.go
index b7702f80c072..7805e8515ccc 100644
--- a/bootstrap/kubeadm/internal/controllers/kubeadmconfig_controller.go
+++ b/bootstrap/kubeadm/internal/controllers/kubeadmconfig_controller.go
@@ -279,7 +279,7 @@ func (r *KubeadmConfigReconciler) Reconcile(ctx context.Context, req ctrl.Reques
 	}
 
 	// Note: can't use IsFalse here because we need to handle the absence of the condition as well as false.
-	if !conditions.IsTrue(cluster, clusterv1.ControlPlaneInitializedCondition) {
+	if !annotations.IsTakeOverCluster(cluster.GetObjectMeta()) && !conditions.IsTrue(cluster, clusterv1.ControlPlaneInitializedCondition) {
 		return r.handleClusterNotInitialized(ctx, scope)
 	}
 
diff --git a/controlplane/kubeadm/internal/controllers/controller.go b/controlplane/kubeadm/internal/controllers/controller.go
index 05d69b52b00b..b3734de5c77c 100644
--- a/controlplane/kubeadm/internal/controllers/controller.go
+++ b/controlplane/kubeadm/internal/controllers/controller.go
@@ -394,6 +394,11 @@ func (r *KubeadmControlPlaneReconciler) reconcile(ctx context.Context, cluster *
 	switch {
 	// We are creating the first replica
 	case numMachines < desiredReplicas && numMachines == 0:
+		if annotations.IsTakeOverCluster(controlPlane.Cluster.GetObjectMeta()) {
+			// Create a new Machine w/ join
+			log.Info("Scaling up control plane for TakeOverCluster", "Desired", desiredReplicas, "Existing", numMachines)
+			return r.scaleUpControlPlane(ctx, cluster, kcp, controlPlane)
+		}
 		// Create new Machine w/ init
 		log.Info("Initializing control plane", "Desired", desiredReplicas, "Existing", numMachines)
 		conditions.MarkFalse(controlPlane.KCP, controlplanev1.AvailableCondition, controlplanev1.WaitingForKubeadmInitReason, clusterv1.ConditionSeverityInfo, "")
@@ -568,7 +573,7 @@ func (r *KubeadmControlPlaneReconciler) reconcileEtcdMembers(ctx context.Context
 	log := ctrl.LoggerFrom(ctx)
 
 	// If etcd is not managed by KCP this is a no-op.
-	if !controlPlane.IsEtcdManaged() {
+	if annotations.IsTakeOverCluster(controlPlane.Cluster.GetObjectMeta()) || !controlPlane.IsEtcdManaged() {
 		return ctrl.Result{}, nil
 	}
 
diff --git a/controlplane/kubeadm/internal/controllers/helpers.go b/controlplane/kubeadm/internal/controllers/helpers.go
index 3355f942409c..40e5f8b9c4f6 100644
--- a/controlplane/kubeadm/internal/controllers/helpers.go
+++ b/controlplane/kubeadm/internal/controllers/helpers.go
@@ -73,9 +73,11 @@ func (r *KubeadmControlPlaneReconciler) reconcileKubeconfig(ctx context.Context,
 	case err != nil:
 		return ctrl.Result{}, errors.Wrap(err, "failed to retrieve kubeconfig Secret")
 	}
-
-	if err := r.adoptKubeconfigSecret(ctx, cluster, configSecret, kcp); err != nil {
-		return ctrl.Result{}, err
+	
+	if util.IsOwnedByObject(configSecret, cluster) && !util.IsControlledBy(configSecret, kcp) {
+		if err := r.adoptKubeconfigSecret(ctx, cluster, configSecret, kcp); err != nil {
+			return ctrl.Result{}, err
+		}
 	}
 
 	// only do rotation on owned secrets
diff --git a/internal/controllers/cluster/cluster_controller_phases.go b/internal/controllers/cluster/cluster_controller_phases.go
index 741a423cb244..06d130c28557 100644
--- a/internal/controllers/cluster/cluster_controller_phases.go
+++ b/internal/controllers/cluster/cluster_controller_phases.go
@@ -245,7 +245,7 @@ func (r *Reconciler) reconcileControlPlane(ctx context.Context, cluster *cluster
 		if err != nil {
 			return ctrl.Result{}, err
 		}
-		if initialized {
+		if annotations.IsTakeOverCluster(cluster.GetObjectMeta()) || initialized {
 			conditions.MarkTrue(cluster, clusterv1.ControlPlaneInitializedCondition)
 		} else {
 			conditions.MarkFalse(cluster, clusterv1.ControlPlaneInitializedCondition, clusterv1.WaitingForControlPlaneProviderInitializedReason, clusterv1.ConditionSeverityInfo, "Waiting for control plane provider to indicate the control plane has been initialized")
diff --git a/util/annotations/helpers.go b/util/annotations/helpers.go
index 0ec9ef9388ac..da4ada4d2c21 100644
--- a/util/annotations/helpers.go
+++ b/util/annotations/helpers.go
@@ -25,6 +25,11 @@ import (
 	clusterv1 "sigs.k8s.io/cluster-api/api/v1beta1"
 )
 
+// IsTakeOverCluster returns true if the object has the `managed-by` annotation.
+func IsTakeOverCluster(o metav1.Object) bool {
+	return hasAnnotation(o, clusterv1.TakeOverCluster)
+}
+
 // IsPaused returns true if the Cluster is paused or the object has the `paused` annotation.
 func IsPaused(cluster *clusterv1.Cluster, o metav1.Object) bool {
 	if cluster.Spec.Paused {