diff --git a/docs/design/feature/distributed_layerwise_offload.md b/docs/design/feature/distributed_layerwise_offload.md
index a52b0b8ea..17fad95d4 100644
--- a/docs/design/feature/distributed_layerwise_offload.md
+++ b/docs/design/feature/distributed_layerwise_offload.md
@@ -56,6 +56,8 @@ AllGather:                [full N+1]             [full N+2]
 Buffers:    [current slot]       [prefetch slot]       [current slot]
 ```
 
+![DLO double-buffer prefetch pipeline](../figures/dlo/dlo_pipeline.gif)
+
 The backend uses two shared device buffers, so accelerator weight residency is
 bounded by the largest streamed blocks rather than the complete model.
 
diff --git a/docs/design/figures/dlo/dlo_pipeline.gif b/docs/design/figures/dlo/dlo_pipeline.gif
new file mode 100644
index 000000000..cc206d03b
Binary files /dev/null and b/docs/design/figures/dlo/dlo_pipeline.gif differ