Did anybody do the math of whether it's more cost effective to pay for all that military to ensure you get cheap silicon made in countries with cheaper labor that in the US rather than just subsidizing the production of semiconductors domestically?
The picture of silicon industry that you paint is very far from reality. Neither Taiwan nor South Korea are third world countries and baking top chips isn't the same as sewing cotton T-shirts or gathering strawberries under the scorching sun. Instead of "cheaper labor", you need top talent and very high quality equipment. That top talent collects very good salaries.
If chip production was just a matter of money, both China and the U.S. would rule the roost. The real bottleneck is talented and loyal engineers.
Semiconductor engineers salaries in Taiwan could've been as low as $26k-$28k USD back in 2007-2009. Multi-year long PhD "interships" can be unpaid, or completely minimally so.
A chance for an average semi process engineer graduate to survive to doing real RnD was close to 80-100 to 1.
I wonder what the situation looks like today. Probably better, because losing senior engineers in a situation when the field has shrunk to Samsung, TSMC and Intel, would be a huge pain.
The mainland is taking advantage of this to get skilled engineers, as mainland Chinese tech salaries are closer to those in the US. So much so that Taiwan has banned recruiting for mainland jobs: https://asia.nikkei.com/Business/Tech/Semiconductors/Taiwan-...
- How much military spending is really what other Western countries would call welfare - not only do you pay the US recruits, they get pensions, they and their families get healthcare etc etc.
- The payback of being the worlds only superpower has been enormous anyway. This is probably a marginal cost.
- It's not entirely clear that building a fab is like dropping a supermarket into place. Whilst in operation they are lights out I sincerely doubt they can be easily replicated - at the very least you will need to just kidnap the whole middle management layer of TSMC and move them to Texas. Which is more or less the current negotiations:-)
In short, yeah the US (and EU) should spend a fortune to build out their own internal chip capacity. Just as we should have done it for oil or electric batteries or ... but we just don't. China however has. Something democracy is good at. Somethings it ain't.
There are more people sitting in offices there than on the floor.
I fact there is a huge oversupply of semiconductor engineers globally because of fabs becoming less, and less labour intensive.
Companies take a big effort to reduce the number of people on the floor futher, but now not so much for labour cost readon, but security. The chance of somebody who shouldn't really be on the floor acidentally pressing the wrong button, and sending $10M worth of wafers down the drain is the risk they don't want to take.
1. Labour-wise probably 3-4 times the Taiwan, but labour cost of fabs is microscopic in comparison to everything else.
Taiwan does not have dramatically lower taxes than US, but there is something particular to how working capital is accounted which will make for a double digit difference in the in TW vs. US.
Third is supplies. Even before the semi moving to Asia, US fabs had to import a big portion of their supplies.
2. > Why did the US outsourced the semiconductor fabrication in the first place
Why did US outsource almost everything, even when it makes no sense?
It simply tough, tedious, problematic doing business in the US.
I wrote about it many times before here.
You mind your own business, very much literally. Few month down the line some trouble comes: lawsuit, taxmen, utilities, creditors, SEC, city hall, suppliers, random activists, labour union... pick any.
What are their home countries? Sorry I could do my own research, but since you made a comment that implied you had a piece of information at hand, I feel I may just get it easier this way.
Exyte (Germany) was a big one. I myself haven't been up to date with news in the industry for around 10 years since I abandoned all attempts to enter the industry.
As for 2), it is less about outsourcing and more about gradual loss of competitiveness. The scientific and technological development between 2000 and 2020 led to a lot of corporations, not just in the U.S., falling out of the race because they could not keep up the pace. Nowadays, only Samsung and TSMC are left, with Intel lagging behind, but still not completely out.
Similar consolidations have happened in the past and in other industrial fields. This time, two of the three surviving champions are simply not American.
This is an interesting take, I'd be curious how it applies to other industries. Is outsourcing an artifact of loss of technical/execution competence required to be competitive?
In general its a dirty process. There are a lot of toxic chemicals involved. People might get sick. Margins aren't as good as software and its very capital intensive so investors would rather invest in software. Lots of booms and bust in the semiconductor industry.
As the US gets more than "cheap labor" for its troop deployments.
Very large parts of it is protecting democracy. Whether you believe that's worth protecting is a different question, but it's left over from cold war & proxy wars against Russia in both Vietnam and Korea.
This is still once way the US exerts its presence against China. Maybe futile, but it is more than just "cheap labor".
Honestly curious about why so many downvotes. It's a honest question. Perhaps you all know the answer and think it's a stupid question not worth answering it. If so please ignore it (other people have answered).
