Hacker News Reader

Comment by Grimburger | Hacker News Reader

I get its sarcasm but TSMC has done so much to try to get the Americans getting some decent yields in their new factories and it's failing so hard, there's real questions about whether they can ever meet their targets. They've sent thousands of existing Taiwanese workers to help at this point and from anecdotal things I hear it's not going well. (I live in Taiwan)

This stuff is all blood, sweat and tears to get it done, and to be honest a culture.

You can't just buy it off the shelf. If that was possible everyone would have gone paid for it like an Ikea product years ago.

View on HN

Investment is risky and uncertain. If you did not do anything because it might fail, you would do nothing at all.

The inability to cope with failure is a culture that is common in Europe. Failure is so personally devastating, so destructive to social and cultural standing that not only must you not take risks but other people must be actively prevented from taking risk.

Maybe the US investment doesn't work, but the US will have tried at least. Taiwan is, ironically, the perfect example of where the initial impetus came from the government because there was no culture of risk-taking. Parts of Europe have a savings culture that is very similar to Taiwan and led to very poor growth and very high inequality.

If the US keeps trying, they will succeed. On here, the view was that China could never build out this capacity, they did. The only way to guarantee failure is to never try.

There are businesses and business failures and government project failures in Europe all the time.

From neglect, not enterprise. The level of failure is also too low.

It’s more about not being able to hire PhDs for $50k a year, tell them it’s a lot of money and make them work Asian hours and efficiency

then just hire them in asia.. and stick to globalization...?

That's part of the culture OP was referencing. Another part is not having ad tech jobs easily available paying 5-10x as much for easier work. A big part of Asian success at hard tech fields is just not having better options.

There are probably zero STEM PhDs working for TSMC making 50K USD total comp per year. The bonuses are enormous at TSMC these days.

The amount of supercars and high end vehicles in Hsinchu is unreal. Probably all anyone needs to know here, until you've seen it firsthand you wouldn't understand.

Great point. I read about this a few times. Here is a juicy quote that I found: "Ferrari sales in Taiwan doubled over a four-year period." Yikes! Hello, new money!

https://www.bloomberg.com/news/articles/2024-10-24/tsmc-s-ar...

Older article, but seems like things are going pretty ok?

They've got huge monetary losses, if you're ok with that sure but its not like any other chip fab on the planet is losing cash.

The Taiwanese are doing a lot to try get it up but all the reports on the ground level are rough.

https://www.tomshardware.com/tech-industry/semiconductors/ts...

> They've got huge monetary losses, if you're ok with that

USA passed the CHIPS Act and is spending close to a trillion dollars to strategically re-home all of TSMC's most advanced fab technology to Arizona as a matter of national security. Monetary losses don't matter when the point is to not be reliant or China in the event of South China Sea conflict.

I think the point of this article, and the warnings from ASML, is that Europe is not making the same effort.

But that's sorta of hits to the heart of my comment, it's not particularly working well. TSMC has 2 fabs in China and yields+costs are a fraction of whats coming out of Arizona, let alone in Hsinchu.

Like are you really going to buy a chip that costs 70% more out of national pride? Might work on some but think that very few else will sign up for that.

Pride has nothing to do with it. It's about being able to get the chips at all when an adversary controls the supply.

Will you buy a computer chip with the same specs that costs 70% more? Please be honest with us all.

You'll do that because of national adversity?

People are spending more than 70% more on memory because that's all that is available. Everything seems to indicate that people will pay a lot more.

70% extra on top of a trillion or two in investment is a relatively small price to pay when faced with the possibility of that being your only option at some point in the future.

I'll admit that I haven't followed the chatter about TSMC Arizona at all. But how does what you say square with https://semiwiki.com/forum/threads/tsmc-overseas-fabs-see-pr...?

I don't want to speak bad about them but it's just general demand for semiconductors at this point, everyone is just buying it all up because they can. The Arizona plant still doesn't seem to be pumping out and the self-sufficiency target for the US doesn't seem all there. Im sorry to be the bearer of bad news but honestly it's a culture that's hard to replicate and its how it somehow gets done in one country on Earth. The 2030 target for the states surely isn't happening, it's way too ambitious.

People really don't understand how this stuff is borderline magic turning silicon into these things. There's so much that can go wrong at every step, it's the most advanced manufacturing process on Earth.

Once upon a time chips went down in value but I don't think that's happening anymore, lot of other people seem aware of that. Not to mention the semiconductor talent is getting eaten up like pancakes everywhere.

Do you have local sources we can look at? I’m not familiar with Taiwan but would be interested reading more about the issues

Reply on news.ycombinator.com