[TMWNN]

>But even according to their flawed analysis, they assumed there would never be a need for so many flights

This is an example of Musk's tech background in action. He knew from his career that if you sell something useful for a very low price, uses for it appear that hadn't existed before. SpaceX's hardware development method of blowing rockets up until one works is basically a edit-compile-run debug cycle.

>and re-use would just eat into their production line productivity, raising the cost of each rocket.

Arianespace specifically and infamously said that reusability would be bad because it would leave rocket assembly crews with nothing to do. I wouldn't be surprised if ULA had the same issue in mind, but at least no one there was dumb enough to publicly say so.