diff --git a/src/huggingface_hub/serialization/_torch.py b/src/huggingface_hub/serialization/_torch.py
index 3c8b2b33f3..e87e728d23 100644
--- a/src/huggingface_hub/serialization/_torch.py
+++ b/src/huggingface_hub/serialization/_torch.py
@@ -368,18 +368,21 @@ def _get_unique_id(tensor: "torch.Tensor") -> Union[int, Tuple[Any, ...]]:
     return unique_id
 
 
-def get_torch_storage_id(tensor: "torch.Tensor") -> Tuple["torch.device", Union[int, Tuple[Any, ...]], int]:
+def get_torch_storage_id(tensor: "torch.Tensor") -> Optional[Tuple["torch.device", Union[int, Tuple[Any, ...]], int]]:
     """
     Return unique identifier to a tensor storage.
 
-    Multiple different tensors can share the same underlying storage. For
-    example, "meta" tensors all share the same storage, and thus their identifier will all be equal. This identifier is
+    Multiple different tensors can share the same underlying storage. This identifier is
     guaranteed to be unique and constant for this tensor's storage during its lifetime. Two tensor storages with
     non-overlapping lifetimes may have the same id.
+    In the case of meta tensors, we return None since we can't tell if they share the same storage.
 
     Taken from https://github.com/huggingface/transformers/blob/1ecf5f7c982d761b4daaa96719d162c324187c64/src/transformers/pytorch_utils.py#L278.
     """
-    return tensor.device, _get_unique_id(tensor), get_torch_storage_size(tensor)
+    if tensor.device.type == "meta":
+        return None
+    else:
+        return tensor.device, _get_unique_id(tensor), get_torch_storage_size(tensor)
 
 
 def get_torch_storage_size(tensor: "torch.Tensor") -> int: