We should not under-estimate the need for speed in supply chains. Predicting future demand is hard. To be more specific, we're talking about predicting ~100M unique products (the order of magnitude that moves on the pacific) and some of them have very lumpy demand (e.g. invent a new product, but it depends on 100 other obscure products).
We should also not over-estimate the need for speed. Just because some items need speed, it does not follow that all items need speed.