diff --git a/go/vt/vttablet/tabletserver/state_manager.go b/go/vt/vttablet/tabletserver/state_manager.go
index 8aa7776957f..9c01610f770 100644
--- a/go/vt/vttablet/tabletserver/state_manager.go
+++ b/go/vt/vttablet/tabletserver/state_manager.go
@@ -99,6 +99,10 @@ type stateManager struct {
 	alsoAllow      []topodatapb.TabletType
 	reason         string
 	transitionErr  error
+	// requestsWaitCounter is the number of goroutines that are waiting for requests to be empty.
+	// If this value is greater than zero, then we have to ensure that we don't Add to the requests
+	// to avoid any panics in the wait.
+	requestsWaitCounter int
 
 	requests sync.WaitGroup
 
@@ -354,6 +358,20 @@ func (sm *stateManager) checkMySQL() {
 	}()
 }
 
+// addRequestsWaitCounter adds to the requestsWaitCounter while being protected by a mutex.
+func (sm *stateManager) addRequestsWaitCounter(val int) {
+	sm.mu.Lock()
+	defer sm.mu.Unlock()
+	sm.requestsWaitCounter += val
+}
+
+// waitForRequestsToBeEmpty waits for requests to be empty. It also increments and decrements the requestsWaitCounter as required.
+func (sm *stateManager) waitForRequestsToBeEmpty() {
+	sm.addRequestsWaitCounter(1)
+	sm.requests.Wait()
+	sm.addRequestsWaitCounter(-1)
+}
+
 func (sm *stateManager) setWantState(stateWanted servingState) {
 	sm.mu.Lock()
 	defer sm.mu.Unlock()
@@ -392,7 +410,9 @@ func (sm *stateManager) StartRequest(ctx context.Context, target *querypb.Target
 	}
 
 	shuttingDown := sm.wantState != StateServing
-	if shuttingDown && !allowOnShutdown {
+	// If requestsWaitCounter is not zero, then there are go-routines blocked on waiting for requests to be empty.
+	// We cannot allow adding to the requests to prevent any panics from happening.
+	if (shuttingDown && !allowOnShutdown) || sm.requestsWaitCounter > 0 {
 		// This specific error string needs to be returned for vtgate buffering to work.
 		return vterrors.New(vtrpcpb.Code_CLUSTER_EVENT, vterrors.ShuttingDown)
 	}
@@ -560,7 +580,7 @@ func (sm *stateManager) unserveCommon() {
 	log.Info("Finished Killing all OLAP queries. Started tracker close")
 	sm.tracker.Close()
 	log.Infof("Finished tracker close. Started wait for requests")
-	sm.requests.Wait()
+	sm.waitForRequestsToBeEmpty()
 	log.Infof("Finished wait for requests. Finished execution of unserveCommon")
 }
 
diff --git a/go/vt/vttablet/tabletserver/state_manager_test.go b/go/vt/vttablet/tabletserver/state_manager_test.go
index 4b88ce734d7..cd72c7232c8 100644
--- a/go/vt/vttablet/tabletserver/state_manager_test.go
+++ b/go/vt/vttablet/tabletserver/state_manager_test.go
@@ -701,6 +701,29 @@ func TestRefreshReplHealthLocked(t *testing.T) {
 	assert.False(t, sm.replHealthy)
 }
 
+// TestPanicInWait tests that we don't panic when we wait for requests if more StartRequest calls come up after we start waiting.
+func TestPanicInWait(t *testing.T) {
+	sm := newTestStateManager(t)
+	sm.wantState = StateServing
+	sm.state = StateServing
+	sm.replHealthy = true
+	ctx := context.Background()
+	// Simulate an Execute RPC running
+	err := sm.StartRequest(ctx, sm.target, false)
+	require.NoError(t, err)
+	go func() {
+		time.Sleep(100 * time.Millisecond)
+		// Simulate the previous RPC finishing after some delay
+		sm.EndRequest()
+		// Simulate a COMMIT call arriving right afterwards
+		_ = sm.StartRequest(ctx, sm.target, true)
+	}()
+
+	// Simulate going to a not serving state and calling unserveCommon that waits on requests.
+	sm.wantState = StateNotServing
+	sm.waitForRequestsToBeEmpty()
+}
+
 func verifySubcomponent(t *testing.T, order int64, component any, state testState) {
 	tos := component.(orderState)
 	assert.Equal(t, order, tos.Order())