diff --git a/vsb/workloads/dataset.py b/vsb/workloads/dataset.py
index fede11d..c9e9451 100644
--- a/vsb/workloads/dataset.py
+++ b/vsb/workloads/dataset.py
@@ -1,11 +1,14 @@
 from collections.abc import Iterator
 
+import gevent
+from gevent.lock import BoundedSemaphore
 import numpy
 import pyarrow.parquet
 from google.cloud.storage import Bucket, Client, transfer_manager
 import json
 import pandas
 import pathlib
+
 from pinecone.grpc import PineconeGRPC
 import pyarrow.dataset as ds
 from pyarrow.parquet import ParquetDataset, ParquetFile
@@ -204,23 +207,31 @@ def should_download(blob):
                 "  ✔ Dataset download complete",
                 total=len(to_download),
             ) as download_task:
-                for blob in to_download:
-                    logger.debug(
-                        f"Dataset file '{blob.name}' not found in cache - will be downloaded"
-                    )
-                    dest_path = self.cache / blob.name
-                    dest_path.parent.mkdir(parents=True, exist_ok=True)
-                    blob.download_to_file(
-                        ProgressIOWrapper(
-                            dest=dest_path,
-                            progress=vsb.progress,
-                            total=blob.size,
-                            scale=1024 * 1024,
-                            indent=2,
+                # Limit the max number of files concurrently downloaded.
+                semaphore = BoundedSemaphore(5)
+
+                def download_file(blob, semaphore):
+                    with semaphore:
+                        logger.debug(
+                            f"Dataset file '{blob.name}' not found in cache - will be downloaded"
                         )
-                    )
-                    if vsb.progress:
-                        vsb.progress.update(download_task, advance=1)
+                        dest_path = self.cache / blob.name
+                        dest_path.parent.mkdir(parents=True, exist_ok=True)
+                        blob.download_to_file(
+                            ProgressIOWrapper(
+                                dest=dest_path,
+                                progress=vsb.progress,
+                                total=blob.size,
+                                scale=1024 * 1024,
+                                indent=2,
+                            )
+                        )
+                        if vsb.progress:
+                            vsb.progress.update(download_task, advance=1)
+
+                greenlets = [gevent.spawn(download_file, blob, semaphore) for blob in to_download]
+                gevent.joinall(greenlets)
+
             # Clear the progress bar now we're done.
             vsb.progress.stop()
             vsb.progress = None