diff --git a/torchsnapshot/io_preparer.py b/torchsnapshot/io_preparer.py
index f837701..e4d9e1f 100644
--- a/torchsnapshot/io_preparer.py
+++ b/torchsnapshot/io_preparer.py
@@ -97,14 +97,7 @@ def prepare_write(
         return entry, []
 
     storage_path = get_storage_path(obj, logical_path, rank, replicated)
-    if isinstance(obj, ShardedTensor):
-        return ShardedTensorIOPreparer.prepare_write(
-            storage_path=storage_path,
-            obj=obj,
-            is_async_snapshot=is_async_snapshot,
-            _tensor_prepare_func=_tensor_prepare_func,
-        )
-    elif isinstance(obj, torch.Tensor):
+    if isinstance(obj, torch.Tensor):
         if obj.nelement() * obj.element_size() > get_max_chunk_size_bytes():
             chunking_instruction = ChunkedTensorIOPreparer.chunk_tensor(obj)
             entry, obj_write_req = ChunkedTensorIOPreparer.prepare_write(