Post

One thing that we didn't seem to talk about when we were talking about waterfalls is that you can more easily cache them on the server. If you have a data fetch dependent on another data fetch on the server, you may be able to cache the first one, or both.

Even if they have to go and do a lookup in dynamic rendering, that may go so fast 99% of the time that you've effectively flattened the waterfall. This is way harder to do with a high hit rate on the client.