perhaps you understand the supply chains / economic factors better?
It seems China is able to make electronics start to finish without importing...
Discussion
lol that’s not true, China imports a ton for their electronics, but more importantly, you can’t start from zero, if you want people to manufacture you either have to dump a few hundred trillion into factories (which we don’t have the manpower or supplies to build, doubly so since they’re now tariff’d lol) or be willing to import low added value goods so that you can build the high added value goods.
I generally agree with the sentiment of your original post.
What I find frustrating is that I've been asking 'how' to localize chipfab in the US for about half a decade, and to my recollection have had zero constructive conversations about it.
People mostly say 'oh, they do really complicated things with big expensive machines....that's about all I know'.
Tariffs come and go no? Did we have budding hardware businesses recently? Would love to dig deeper into the space.
Chipfabs specifically is a question of being bleeding edge. There are a lot of chipfabs (including some post-fab processing) in the US (they’re high value-added goods, we manufacture a lot of those in the US!). But it’s fierce competition between Intel and TSMC. So to get Intel to be dominant it’s mostly a question of fixing bad leadership at Intel and making sure the best engineers want to work there.
That seems incredibly unlikely. Do you see any other paths?
Roughly - it seems a combination of having the right city infrastructure (and style), in combination with a labor force that can and will do the work.
This is my working theory after some light research on Tapei/TSMC.
If I remember correctly - TSMC was going to manufacture here but didn't see the workforce for it. I'd imagine there are other factors in play, but that one did seem pretty real to me.
I didn’t say it was likely, said it was the only path to having a fully domestic chip fab. TSMC absolutely manufactures a good bit in the US now, but never the absolute latest generation, and for political reasons isn’t likely to change that.
Going on the assumption that one wants to improve things (and that paying too much attention to tsmc, or intel is a waste of time).
It's still pretty unclear what any sort of grass roots approach could/would be.
I think we can do better than....what? "Start with getting an electrical engineering degree".
You absolutely cannot start from scratch here. I mean if you have like a trillion dollars to burn hiring engineers, buying equipment, and creating process nodes that you throw away maybe? But even then you’re probably better off just acquiring Intel and then driving their process nodes to leadership.
I'm tempted to disagree, but I'm also fairly out of my depth.
Putting the idea of a company aside for a second though...the path to learning the skills could be made more accessible, or do you think not?