FYI - if anyone is curious to see HSDS used with really large datasets (> 1TB), checkout this repo:
GitHub - NREL/hsds-examples: Examples of using the HSDS Service to Access NREL WIND Toolkit data. It’s hosted by NREL, but contains many Python notebooks that illustrate using h5pyd with HSDS to access data released by NREL and hosted on AWS S3.
The repo supports Github codespaces, so you can just run the examples from your browser (no setup needed). To get started, click the codespaces link in the README.
This AWS blog post came out several years ago, but does a good job of explaining the benefits of using HSDS accessing the wind datasets: Power from wind: Open data on AWS | AWS Big Data Blog. Since this was publish, the amount of HDF data published by NREL has increased greatly and is now several PBs.