That's presuming that such a conflict would be fought as the sort of industrial war that WW1 and WW2 were. But there's little evidence that's the case: the lead

That's presuming that such a conflict would be fought as the sort of industrial war that WW1 and WW2 were. But there's little evidence that's the case: the lead times on modern weapons systems are enormous, and the attrition rate of hardware is enormous too (i.e. consider that the Javelin can put a tank out of action permanently). It is an unproven criteria that it is reasonable or possible to expect to produce the munitions and equipment for a conflict during the conflict if it's the sort of near-peer thing people usually cite for this - i.e. rather you go to war with what you have, and you're unlikely to rebuild that advantage in time if you don't win then. Which is the point: localizing a bunch of industries on that basis may not make any sense, compared to simply stockpiling from the cheapest accreditable sources.