←back to thread

TSMC looks to double down on U.S. chip factories as talks in Europe falter

(www.reuters.com)
437 points adventured | 4 comments | 15 May 21 00:49 UTC | HN request time: 0.723s | source
Show context
SCAQTony ◴[15 May 21 02:16 UTC] No.27161687[source]▶
The USA is also willing to defend Taiwan and currently the US has carrier groups in the south China sea. This suggests that the idea of building chips in the US may be a diplomatic courtesy and arguably an incentive.
replies(7): >>27161727 #>>27161793 #>>27161931 #>>27161972 #>>27161998 #>>27162036 #>>27162628 #
crocodiletears ◴[15 May 21 03:16 UTC] No.27161998[source]▶
I would argue that by the time China comes to take Taiwan (when, not if), the US will be in a position where it is unwilling/able to expend the lives and funds required to provide a prolonged defense of a piece of land firmly in China's front yard.

Personally, I think we're at the point where the US would consider the costs of Taiwan's defense well beyond any potential benefits. It's too economically integrated with China, and the sheer number of bodies it would take aren't worth the moral victory.

Hong Kong is the writing on the wall. China wishes to restore integrity to what it regards as its territory.

TSMC must open facilities in the west because its Taiwan facilities are too dangerous to leave in enemy hands. It has to invest in capital outside of any potential conflict zone if it plans to exists over the long-term as a profit-making entity.

replies(8): >>27162099 #>>27162114 #>>27162121 #>>27162124 #>>27162210 #>>27162415 #>>27162470 #>>27164467 #
1. systemvoltage ◴[15 May 21 03:38 UTC] No.27162114[source]▶
Our memory is fading. Just 5-6 years ago, Intel fabs had the best process technology and TSMC was behind. No one could have guessed what the state of the world would be in 2021 with any significant certainty. Who knows what would happen in another 5 years? We are so myopic when it comes to the bias towards the status quo. We give more weight to what the current situation is and then extrapolate that the future will be exactly the same. We are bad at projecting the future.

Is it worth a war that can last many years? By the end of the war, what if US doesn’t need TSMC technology and the tides are reversed?

I’m not conjecturing, I’m shedding light on how much we are willing to risk now vs a few years later. It could also be that Intel fabs fall further behind and TSMC becomes 10x more important than now. Who knows!!!

Instead of war, invest in domestic technology. It pays back better.

replies(2): >>27162222 #>>27164785 #
2. totalZero ◴[15 May 21 03:59 UTC] No.27162222[source]▶
Your last sentence is exactly on point. What's sad is that we spend far more on naval power projection than on detours around obvious risks to prosperity and peace, such as the semiconductor supply chain. We are barely spending 50 billion dollars on domestic semiconductor fabrication in the upcoming legislation, despite acknowledging that the economic costs of the semiconductor shortage are far greater. Contrast that 50 billion number with Apple's expenditures on buybacks and dividends over the past decade. They aren't even in the same order of magnitude.
replies(1): >>27162338 #
3. systemvoltage ◴[15 May 21 04:21 UTC] No.27162338[source]▶
Indeed, semiconductor industry investment is a drop in a bucket compared to 10 years of war. Not to mention, after a war, there still exists a dependency.

It's funny how most of the semiconductor topics lead to war discussions but no one talks about outcompeting Taiwan. It's doable if there is will and resources (I am a Fab engineer). IMO the best way to move forward.

4. ta_ca ◴[15 May 21 12:58 UTC] No.27164785[source]▶
> Instead of war, invest in domestic technology. It pays back better

is it though? if you got the biggest gun why not rob someone else? do you believe ethics/empathy/humanity plays any role in any form governance? since we know that if you can sell the idea, you can get away with anything. why not just do the thing you know best? this is not singling out USA or China in any sense, they are just ahead of the curve, flower of the century.

a system optimized for only profit with almost no accountability (externalities) reads like a recipe to transfer of the wealth from many to few. if you add war into equation many becomes many-brown, many-yellow, many-black, many-jewish, many-others basicaly many-we-can-target-and-get-away-with. if you remove borders from the equation you end up seeing yourself on the receiving end as well. yet no-surprisingly few remains always the same.

so, the question is 'it pays back better to whom?'.