diff --git a/service/storage_service.cc b/service/storage_service.cc
index 21357c43c0..534ee77c2d 100644
--- a/service/storage_service.cc
+++ b/service/storage_service.cc
@@ -1834,9 +1834,6 @@ future<> storage_service::join_topology(sharded<db::system_distributed_keyspace>
             }
         }
 
-        // If we were the first node in the cluster, at this point `upgrade_state` will be
-        // initialized properly. Yield control to group 0
-        _manage_topology_change_kind_from_group0 = true;
         set_topology_change_kind(upgrade_state_to_topology_op_kind(_topology_state_machine._topology.upgrade_state));
 
         co_await update_topology_with_local_metadata(*raft_server);
@@ -1870,7 +1867,6 @@ future<> storage_service::join_topology(sharded<db::system_distributed_keyspace>
         co_return;
     }
 
-    _manage_topology_change_kind_from_group0 = true;
     set_topology_change_kind(upgrade_state_to_topology_op_kind(_topology_state_machine._topology.upgrade_state));
 
     // We bootstrap if we haven't successfully bootstrapped before, as long as we are not a seed.
@@ -2948,8 +2944,7 @@ future<> storage_service::join_cluster(sharded<db::system_distributed_keyspace>&
         slogger.info("Raft recovery - starting in legacy topology operations mode");
         set_topology_change_kind(topology_change_kind::legacy);
     } else if (_group0->joined_group0()) {
-        // We are a part of group 0. The _topology_change_kind_enabled flag is maintained from there.
-        _manage_topology_change_kind_from_group0 = true;
+        // We are a part of group 0.
         set_topology_change_kind(upgrade_state_to_topology_op_kind(_topology_state_machine._topology.upgrade_state));
         if (_db.local().get_config().force_gossip_topology_changes() && raft_topology_change_enabled()) {
             throw std::runtime_error("Cannot force gossip topology changes - the cluster is using raft-based topology");
diff --git a/service/storage_service.hh b/service/storage_service.hh
index 5e1f034fbe..75457292c8 100644
--- a/service/storage_service.hh
+++ b/service/storage_service.hh
@@ -809,13 +809,6 @@ private:
     // After the node successfully joins, the control over the variable is yielded
     // to `topology_state_load`, so that it can control it during the upgrade from gossiper
     // based topology to raft-based topology.
-    // FIXME: This boolean flag is mostly needed because, shortly after starting
-    // a new group 0, the state of `system.topology` is empty, and updating the
-    // `_topology_change_kind_enabled` variable from group 0 state would lead
-    // to wrong results. This could be fixed by writing an initial `system.topology`
-    // state before creating the group 0 (also making sure to set the correct
-    // group 0 state id so that the timestamps of further mutations are correct).
-    bool _manage_topology_change_kind_from_group0 = false;
     topology_change_kind _topology_change_kind_enabled = topology_change_kind::unknown;
 
     // Throws an exception if the node is either starting and didn't determine which