diff --git a/src/lerobot/datasets/streaming_dataset.py b/src/lerobot/datasets/streaming_dataset.py
index e02550e2f..b63e34dc8 100644
--- a/src/lerobot/datasets/streaming_dataset.py
+++ b/src/lerobot/datasets/streaming_dataset.py
@@ -58,6 +58,10 @@ class LookAheadError(Exception):
     pass
 
 
+class _ShardExhaustedError(Exception):
+    """Raised when a streaming dataset shard has no more items."""
+
+
 class Backtrackable[T]:
     """
     Wrap any iterator/iterable so you can step back up to `history` items
@@ -422,10 +426,7 @@ class StreamingLeRobotDataset(torch.utils.data.IterableDataset):
                     else:
                         frames_buffer.append(frame)
                     break  # random shard sampled, switch shard
-            except (
-                RuntimeError,
-                StopIteration,
-            ):  # NOTE: StopIteration inside a generator throws a RuntimeError since python 3.7
+            except _ShardExhaustedError:
                 del idx_to_backtrack_dataset[shard_key]  # Remove exhausted shard, onto another shard
 
         # Once shards are all exhausted, shuffle the buffer and yield the remaining frames
@@ -503,7 +504,11 @@ class StreamingLeRobotDataset(torch.utils.data.IterableDataset):
 
     def make_frame(self, dataset_iterator: Backtrackable) -> Generator:
         """Makes a frame starting from a dataset iterator"""
-        item = next(dataset_iterator)
+        try:
+            item = next(dataset_iterator)
+        except StopIteration as e:
+            # Translate exhaustion here, before PEP 479 turns it into an indistinguishable RuntimeError.
+            raise _ShardExhaustedError from e
         item = item_to_torch(item)
 
         updates = []  # list of "updates" to apply to the item retrieved from hf_dataset (w/o camera features)
diff --git a/tests/datasets/test_streaming.py b/tests/datasets/test_streaming.py
index cae4be5b6..08544326f 100644
--- a/tests/datasets/test_streaming.py
+++ b/tests/datasets/test_streaming.py
@@ -252,6 +252,54 @@ def test_frames_order_with_shards(tmp_path, lerobot_dataset_factory, shuffle):
             assert frames_match
 
 
+def test_iter_raises_on_frame_error(tmp_path, lerobot_dataset_factory, monkeypatch):
+    """Video decode failures must propagate instead of being silently ignored."""
+    ds_num_frames = 20
+    ds_num_episodes = 2
+    buffer_size = 10
+
+    local_path = tmp_path / "test"
+    repo_id = f"{DUMMY_REPO_ID}"
+
+    lerobot_dataset_factory(
+        root=local_path,
+        repo_id=repo_id,
+        total_episodes=ds_num_episodes,
+        total_frames=ds_num_frames,
+    )
+
+    streaming_ds = StreamingLeRobotDataset(repo_id=repo_id, root=local_path, buffer_size=buffer_size)
+
+    def broken_video_decode(*args, **kwargs):
+        raise RuntimeError("Could not load libtorchcodec")
+
+    monkeypatch.setattr(streaming_dataset_module, "decode_video_frames_torchcodec", broken_video_decode)
+
+    with pytest.raises(RuntimeError, match="libtorchcodec"):
+        next(iter(streaming_ds))
+
+
+def test_iter_raises_on_nested_generator_error(tmp_path, lerobot_dataset_factory, monkeypatch):
+    """PEP 479 errors below frame construction must not be mistaken for shard exhaustion."""
+    local_path = tmp_path / "test"
+    repo_id = DUMMY_REPO_ID
+
+    lerobot_dataset_factory(root=local_path, repo_id=repo_id, total_episodes=2, total_frames=20)
+    streaming_ds = StreamingLeRobotDataset(repo_id=repo_id, root=local_path, buffer_size=10)
+
+    def broken_frame_generator():
+        raise StopIteration("decoder internal failure")
+        yield
+
+    def broken_video_decode(*args, **kwargs):
+        return next(broken_frame_generator())
+
+    monkeypatch.setattr(streaming_dataset_module, "decode_video_frames_torchcodec", broken_video_decode)
+
+    with pytest.raises(RuntimeError, match="generator raised StopIteration"):
+        next(iter(streaming_ds))
+
+
 @pytest.mark.parametrize(
     "state_deltas, action_deltas",
     [