I'm skeptical that peacetime military-dedicated manufacturing capability + peacetime stockpiles would be able to scale to meet wartime needs. I would think it would be significantly advantageous to have plenty of slack, state-of-the-art civilian capacity (+ related expertise) in all needed areas.
At 1.3 million active duty military and more than 800,000 reserve forces the infrastructure to maintain that is already significant manufacturing capacity. Training more people takes time so you get various projections around force mobilization vs war footing etc. But while some lag is expected, you can get a lot done with even 1 year of critical supplies on hand.
Remember this is the same group that designed the internet to keep working even as major cities where nuked. Going to war is very much part of long term planning.
Of note, the minimal types of equipment, large existing stockpiles, and lower percentage of servicemen vs civilians even in a total war situation makes this a vastly smaller market than consumer electronics. The risk vs benifits of offshoring is simply not worth it. Especially becase it allows congressional spending in specific districts.
PS: While manufacturing would need to be expanded, stockpiles allows for a significant ramp up period.