diff --git a/cluster-autoscaler/FAQ.md b/cluster-autoscaler/FAQ.md
index 6eef6433d797..68d0f176a0df 100644
--- a/cluster-autoscaler/FAQ.md
+++ b/cluster-autoscaler/FAQ.md
@@ -882,6 +882,7 @@ The following startup parameters are supported for cluster autoscaler:
 | `balance-similar-node-groups` | Detect similar node groups and balance the number of nodes between them | false
 | `balancing-ignore-label` | Define a node label that should be ignored when considering node group similarity. One label per flag occurrence. | ""
 | `balancing-label` | Define a node label to use when comparing node group similarity. If set, all other comparison logic is disabled, and only labels are considered when comparing groups. One label per flag occurrence. | ""
+| `skip-similar-node-group-recomputation` | Should CA skip similar NodeGroup recomputation for the best option returned by the expander during scaleups. You must enable `balance-similar-node-groups` for this to work. | false
 | `node-autoprovisioning-enabled` | Should CA autoprovision node groups when needed | false
 | `max-autoprovisioned-node-group-count` | The maximum number of autoprovisioned groups in the cluster | 15
 | `unremovable-node-recheck-timeout` | The timeout before we check again a node that couldn't be removed before | 5 minutes
diff --git a/cluster-autoscaler/config/autoscaling_options.go b/cluster-autoscaler/config/autoscaling_options.go
index e9bafdbd8903..2554f8b1d29f 100644
--- a/cluster-autoscaler/config/autoscaling_options.go
+++ b/cluster-autoscaler/config/autoscaling_options.go
@@ -120,6 +120,8 @@ type AutoscalingOptions struct {
 	GpuTotal []GpuLimits
 	// NodeGroupAutoDiscovery represents one or more definition(s) of node group auto-discovery
 	NodeGroupAutoDiscovery []string
+	// SkipSimilarNodeGroupRecomputation skips similar NodeGroup recomputation for the best option returned by the expander during scaleups
+	SkipSimilarNodeGroupRecomputation bool
 	// EstimatorName is the estimator used to estimate the number of needed nodes in scale up.
 	EstimatorName string
 	// ExpanderNames sets the chain of node group expanders to be used in scale up
diff --git a/cluster-autoscaler/core/scaleup/orchestrator/orchestrator.go b/cluster-autoscaler/core/scaleup/orchestrator/orchestrator.go
index 6c6b229bcc10..f85138d5a34d 100644
--- a/cluster-autoscaler/core/scaleup/orchestrator/orchestrator.go
+++ b/cluster-autoscaler/core/scaleup/orchestrator/orchestrator.go
@@ -229,7 +229,7 @@ func (o *ScaleUpOrchestrator) ScaleUp(
 		}
 	}
 
-	scaleUpInfos, aErr := o.balanceScaleUps(now, bestOption.NodeGroup, newNodes, nodeInfos, schedulablePodGroups)
+	scaleUpInfos, aErr := o.balanceScaleUps(now, bestOption, newNodes, nodeInfos, schedulablePodGroups)
 	if aErr != nil {
 		return status.UpdateScaleUpError(
 			&status.ScaleUpStatus{CreateNodeGroupResults: createNodeGroupResults, PodsTriggeredScaleUp: bestOption.Pods},
@@ -680,13 +680,16 @@ func (o *ScaleUpOrchestrator) GetCappedNewNodeCount(newNodeCount, currentNodeCou
 
 func (o *ScaleUpOrchestrator) balanceScaleUps(
 	now time.Time,
-	nodeGroup cloudprovider.NodeGroup,
+	bestOption *expander.Option,
 	newNodes int,
 	nodeInfos map[string]*schedulerframework.NodeInfo,
 	schedulablePodGroups map[string][]estimator.PodEquivalenceGroup,
 ) ([]nodegroupset.ScaleUpInfo, errors.AutoscalerError) {
-	// Recompute similar node groups in case they need to be updated
-	similarNodeGroups := o.ComputeSimilarNodeGroups(nodeGroup, nodeInfos, schedulablePodGroups, now)
+	similarNodeGroups := bestOption.SimilarNodeGroups
+	if !o.autoscalingContext.SkipSimilarNodeGroupRecomputation {
+		// Recompute similar node groups in case they need to be updated
+		similarNodeGroups = o.ComputeSimilarNodeGroups(bestOption.NodeGroup, nodeInfos, schedulablePodGroups, now)
+	}
 	if similarNodeGroups != nil {
 		// if similar node groups are found, log about them
 		similarNodeGroupIds := make([]string, 0)
@@ -699,7 +702,7 @@ func (o *ScaleUpOrchestrator) balanceScaleUps(
 		klog.V(2).Info("No similar node groups found")
 	}
 
-	targetNodeGroups := []cloudprovider.NodeGroup{nodeGroup}
+	targetNodeGroups := []cloudprovider.NodeGroup{bestOption.NodeGroup}
 	for _, ng := range similarNodeGroups {
 		targetNodeGroups = append(targetNodeGroups, ng)
 	}
diff --git a/cluster-autoscaler/main.go b/cluster-autoscaler/main.go
index 71df2f221df9..f195cd48dcae 100644
--- a/cluster-autoscaler/main.go
+++ b/cluster-autoscaler/main.go
@@ -189,6 +189,8 @@ var (
 			"Azure matches by VMSS tags, similar to AWS. And you can optionally specify a default min and max size, e.g. `label:tag=tagKey,anotherTagKey=bar,min=0,max=600`. "+
 			"Can be used multiple times.")
 
+	skipSimilarNodeGroupRecomputation = flag.Bool("skip-similar-node-group-recomputation", false, "if true, skips similar NodeGroup recomputation for the best option returned by the expander during scaleups. You must enable `balance-similar-node-groups` for this to work.")
+
 	estimatorFlag = flag.String("estimator", estimator.BinpackingEstimatorName,
 		"Type of resource estimator to be used in scale up. Available values: ["+strings.Join(estimator.AvailableEstimators, ",")+"]")
 
@@ -445,6 +447,7 @@ func createAutoscalingOptions() config.AutoscalingOptions {
 			MaxAllocatableDifferenceRatio:    *maxAllocatableDifferenceRatio,
 			MaxFreeDifferenceRatio:           *maxFreeDifferenceRatio,
 		},
+		SkipSimilarNodeGroupRecomputation:       *skipSimilarNodeGroupRecomputation,
 		DynamicNodeDeleteDelayAfterTaintEnabled: *dynamicNodeDeleteDelayAfterTaintEnabled,
 		BypassedSchedulers:                      scheduler_util.GetBypassedSchedulersMap(*bypassedSchedulers),
 		ProvisioningRequestEnabled:              *provisioningRequestsEnabled,