diff --git a/torchrec/distributed/batched_embedding_kernel.py b/torchrec/distributed/batched_embedding_kernel.py
index 968d7aa15..2919f4684 100644
--- a/torchrec/distributed/batched_embedding_kernel.py
+++ b/torchrec/distributed/batched_embedding_kernel.py
@@ -246,6 +246,12 @@ def _populate_ssd_tbe_params(config: GroupedEmbeddingConfig) -> Dict[str, Any]:
         ssd_tbe_params["kvzch_eviction_trigger_mode"] = fused_params.get(
             "kvzch_eviction_trigger_mode"
         )
+    if "enable_optimizer_offloading" in fused_params:
+        ssd_tbe_params["enable_optimizer_offloading"] = fused_params.get(
+            "enable_optimizer_offloading"
+        )
+    else:
+        ssd_tbe_params["enable_optimizer_offloading"] = False
 
     ssd_tbe_params["table_names"] = [table.name for table in config.embedding_tables]