diff --git a/ir_datasets/util/download.py b/ir_datasets/util/download.py
index 94d0ccce..dd2dc202 100644
--- a/ir_datasets/util/download.py
+++ b/ir_datasets/util/download.py
@@ -1,6 +1,7 @@
 import json
 import pkgutil
 import os
+import sys
 from pathlib import Path
 import atexit
 from collections import deque
@@ -20,6 +21,20 @@ class BaseDownload:
     def stream(self):
         raise NotImplementedError()
 
+class GoogleCloudBucketStream(BaseDownload):
+    def __init__(self, url, tries=None):
+        self.uri = url.replace("https://storage.googleapis.com/", "gs://")
+        self.tries = tries
+
+    def __repr__(self):
+        return f'GoogleCloudBucketStream({repr(self.uri)}, tries={self.tries})'
+    
+    @contextlib.contextmanager
+    def stream(self):
+        import subprocess
+        proc = subprocess.Popen(['gsutil', 'cat', self.uri], stdout=subprocess.PIPE)
+        with io.BufferedReader(proc.stdout, buffer_size=io.DEFAULT_BUFFER_SIZE) as stream:
+            yield stream
 
 class GoogleDriveDownload(BaseDownload):
     def __init__(self, url, tries=None):
@@ -333,7 +348,9 @@ def __getitem__(self, key):
                 local_msg = (f'If you have a local copy of {dlc["url"]}, you can symlink it here '
                              f'to avoid downloading it again: {local_path}')
                 sources.append(LocalDownload(local_path, local_msg, mkdir=False))
-            if dlc['url'].startswith('https://drive.google.com/'):
+            if dlc['url'].startswith("https://storage.googleapis.com/") and 'google.colab' in sys.modules:
+                sources.append(GoogleCloudBucketStream(dlc['url'], **download_args))
+            elif dlc['url'].startswith('https://drive.google.com/'):
                 sources.append(GoogleDriveDownload(dlc['url'], **download_args))
             else:
                 sources.append(RequestsDownload(dlc['url'], **download_args))