There is of course a large labour cost delta and probably a skills shortage in the US that might be an obstacle for making something like this be economically sustainable/competitive - but if it could be automated sufficiently so that labour cost is less of an issue, is this even feasible in theory?
I know there are some fabs being built, but beyond that is some kind of broader electronics manufacturing hub in the US conceivable?
From the article it sounds like the manufacturing is staying in Asia but moving to politically allied countries where new infra is being built for this. Is it still mainly a labour cost issue?
My complaint is that domestic board houses are super expensive, so I meant competitive with Chinese production. The person replying to me above said they are "marginally more expensive" but another commenter here went through the full quoting process and got a price that is about 8x higher than a reputable fab in China, which is not "marginally more expensive".
So why go domestic? For us, there were IP concerns (that we'd already seen with non-electronic products), there were concerns about shipments getting held up in customs, and there were concerns about manufacturing support. If you want boards to be flashed and tested at the factory, so you can dropship your assembled product to whatever its final destination, its very valuable to send somebody to the board house to work with their integration engineers to do it. I ended up spending most of a week commuting to a board house in Fremont in 2021, so I could train up their people on operating the programming and testing jig we'd built, and operating the calibration system we'd built. Driving from my home (at the time) in Oakland to Fremont was much much easier than flying to Shenzen (or wherever), I didn't have to contend with the great firewall, and if something went wrong, I could be over in 45 minutes. All during the California (and also China) Covid lockdowns.
That is what you're getting with a domestic board house, so if none of that is a concern, there's really no reason to pay the borderline extortionate premium to do it.
And I’m not saying “every metro should have competitive world class manufacturing of every type”, which is such a strong claim as to feel strawman-ish (but I get the sense you didn’t mean to do so). I do however deeply wish we had Shenzhen-like manufacturing hubs on the east and west coasts.