> The team working on it seems to be more interested in the theory than the practice of sharding

As it should be for something that will need to be unbreakable and ultra-reliable. I detect a tone of distaste for theory in your phrasing, which isn't promising.