diff --git a/corelib/dynamicemb/dynamicemb/batched_dynamicemb_tables.py b/corelib/dynamicemb/dynamicemb/batched_dynamicemb_tables.py
index 282f6165d..9770bff77 100644
--- a/corelib/dynamicemb/dynamicemb/batched_dynamicemb_tables.py
+++ b/corelib/dynamicemb/dynamicemb/batched_dynamicemb_tables.py
@@ -1207,6 +1207,11 @@ def export_keys_values(
                     f"Rank {dist.get_rank()} has accumulated count {accumulated_counts} which is different from expected {local_max_rows}, "
                     f"difference: {accumulated_counts - local_max_rows}"
                 )
+
+        if len(keys_list) == 0:
+            return torch.empty(0, dtype=torch.int64, device=device), torch.empty(
+                0, 0, device=device
+            )
         return torch.cat(keys_list), torch.cat(values_list, dim=0)
 
     def incremental_dump(
diff --git a/corelib/dynamicemb/test/test_batched_dynamic_embedding_tables_v2.py b/corelib/dynamicemb/test/test_batched_dynamic_embedding_tables_v2.py
index 22eae8737..6aaaa57ef 100644
--- a/corelib/dynamicemb/test/test_batched_dynamic_embedding_tables_v2.py
+++ b/corelib/dynamicemb/test/test_batched_dynamic_embedding_tables_v2.py
@@ -1143,3 +1143,36 @@ def test_empty_batch(opt_type, opt_params, dim, caching, deterministic, PS):
         del os.environ["DEMB_DETERMINISM_MODE"]
 
     print("all check passed")
+
+
+def test_export_keys_values_empty_table():
+    """export_keys_values() on a never-used table must return empty tensors
+    (not crash on torch.cat([])) -- covers the empty keys_list guard."""
+    assert torch.cuda.is_available()
+    device = torch.device("cuda:0")
+
+    opt = DynamicEmbTableOptions(
+        dim=8,
+        init_capacity=1024,
+        max_capacity=1024,
+        index_type=torch.int64,
+        embedding_dtype=torch.float32,
+        device_id=0,
+        score_strategy=DynamicEmbScoreStrategy.TIMESTAMP,
+        caching=False,
+        local_hbm_for_values=1024**3,
+    )
+    bdebt = BatchedDynamicEmbeddingTablesV2(
+        table_names=["t0"],
+        table_options=[opt],
+        feature_table_map=[0],
+        pooling_mode=DynamicEmbPoolingMode.SUM,
+        optimizer=EmbOptimType.SGD,
+        learning_rate=0.1,
+    )
+
+    keys, values = bdebt.export_keys_values("t0", device)
+
+    assert keys.shape == (0,)
+    assert keys.dtype == torch.int64
+    assert values.shape[0] == 0