I'm in the industry. People have indeed, as you say, been throwing around "re-shoring" and "onshoring" for quite a few years. I've read the articles.
I and many of my colleagues are just wondering where it is. PCB manufacturing may not be as sexy as chip manufacturing, but it's critical.
People really need to be thinking about the constraints the customers are working under. If you are Apple, you can buy from anywhere - you rarely have restrictions. But if you are doing anything under ITAR, where and who makes that matters. It may not be enough for TSMC to have a plant in Arizona, they may need to spin that off into a US held subsidiary. The US military imposes a lot of restrictions on where and who manufactures their components, but they also aren't customers for 2nm, at least for a while. Maybe the US military is shifting their expectations for more cutting edge compute in new platforms (you would imagine their 6th generation fighter concepts would have substantially more advanced compute requirements, which only Intel was really in position to provide, and you can't blame the DOD if they were a little uncertain if that was going to be the case - Intel has has their struggles).
It seems pretty clear to me the target audience for CHIPS, in the end, is the DOD. And rebuilding the US talent pool is part of that. That said, there is still no strategic effort for higher education in this country - I used to build the kinds of programs that the US needs to staff these facilities and there is no strategic effort driving that whatsoever - there's no grants for PhD semiconductor engineer or AA level technician programs or any of the other disciplines involved - industrial engineering, etc. so a lot more is needed to actually get there. So when TSMC says they need to import workers - yeah, part of my job was providing reports to the CA governors office regarding the potential for the state to meet prospective new industry needs for specific technical roles, and if it was anything at Apple's scale or at TSMCs level of expertise and involved
manufacturing, the answer was almost always 'not only do we not have it, our baseline data says the entire country doesn't have it'. If it's design, etc. we're usually great. But we suck at making stuff and we suck at training people to make stuff.
But DOD isn't the only potential customer. NASA wants to go back to the moon, all their shit usually needs to be made in the US (Japan is going to make us a nice Toyota rover, though. I assume Canadas arms were all made in Canada), the new EV subsidies require half the car be made in the US - sure. California throws off a lot of money for industry that needs to be spent in California (why does every EV maker have design offices in the state), etc. How much of this is chasing cutting edge nodes, and how much of this isn't going to have a 2nm design until 2030 (at least). I think the folks arguing that 'hey, this doesn't change where Apple buys their chips' are missing the actual point of this bill. Sure, it'd be nice if it did that, but that's not the primary goal of the bill - the primary goal is to ensure that these 'must be made in the US' things can be made in the US. Having Intel as the sole supplier for any advanced process is eventually (if not already) going to run into a 'having Boeing as the sole supplier for military support aircraft' problem which doesn't look great right now.
The PCB issue is in a lot of ways more important because there's a lot of what the US wants to buy domestically that doesn't need cutting edge node, but does need modern, efficient PCBs. My son is also in this industry (he designs, and they manufacture locally control systems for the semiconductor industry and PCB capability is a real problem for them. They need a 12 layer board that can operate in this temperature range, and due to vibration issues these components need to not just be soldered but glued in place using an adhesive that can handle those temperatures, oh and this is a high radiation environment so here's a pile of other constraints we have on manufacturing - do you have the equipment and engineering experience with that kind of thing, blah, blah, blah. At least they aren't high volume, unlike say Apple. Can they go outside the US? Not usually. And they struggle. They've struggled with domestic suppliers using counterfeit parts, and struggled with domestic suppliers not modernizing. Sometimes you have a whole other set of reasons for reducing your potential pool of contractors.
But I know there's this perception that the US has the greatest higher education system in the world, and all that, and it's excellent - no question, but it's also aimless. The main reason that these plants are going down in Arizona is because ASU throws off a lot of the engineers these plants will need - but that wasn't planned, it was coincidental. There's a good education/industry synergy there. Apple has complained for well over a decade, since back when Obama asked Jobs why Apple couldn't make more of their stuff in the US, that the US didn't produce the kind and volume of engineers needed for these industries. And he's absolutely right. For years we wanted to build an industrial engineering program, but they are expensive to build and to a university, an industrial engineering degree and a sociology degree produces the same amount of tuition, secures the same amount of state tuition subsidy, and the former costs 3x as much as the latter, and 50x as much up front to build the program. Universities, left to their own devices, are never going to build that program because it makes no financial sense to do it, not until the program is sufficiently mature to start pulling in a lot of grant money - and that will take decades. If the country wants these kinds of programs - and universities are absolutely wiling to build them - there needs to be grant money earmarked to build them and after a few years they'll become self-sustaining. There's a huge structural employment problem around manufacturing in general and around high-tech manufacturing particularly that are slow and expensive to build. It's hard to get an industry that doesn't yet exist to fund that pipeline for the 10 years from hiring faculty to producing your first PhD. They're going to be reliant on H1B for a LONG time, which it also turns out becomes an excuse for why the public or some parts of government will oppose the effort entirely. You see that in these forums all the time. In the 14 years since that interaction between Jobs and Obama, a 2nd Obama term, Trump, Biden, there's been no action on that front. And Jobs was hardly the only one saying that. Congress hasn't acted either. We did this back during the Cold War - in fact, that effort played a sizable role in why CA tipped up so many public universities during that period. From 1958 to 1965, UC opened 7 campuses. The CA aerospace industry built off the back of those campuses and the tech industry after that. That was what the federal government wanted to happen, they put the money and focus in that specific direction, and it happened.