> The likely aggressors would lose access to chips to fuel their economies

Economic arguments for war being irrational and therefore unlikely have a history of falling flat on their face. Most infamously, this claim was made about major wars in Europe right before the First World War: https://en.wikipedia.org/wiki/The_Great_Illusion

Unless you're very careful and hedge your words, claiming that X ("war will cause economic harm to both parties") will cause a reduction in Y ("likelihood of war") leads everyone to say your claim is "disproven" whenever Y happens. Doesn't automatically mean the claim has actually been disproven.

It’s a different scenario imo. The economic ties between the US and China are much stronger than the European powers prewar, which were all colonial powers with protectionism and merchantilist trade policy.

News coverage of China is asinine, vacillating between the WSJ/Gordon Chang “the communists are about to collapse” nonsense and the “General <x> says the Chinese are aggressively aggressive and catching up” patter. The chip bullshit is all about Intel lobbying for billions to pull its nuts out of the fire.

IMO Chinese policy is methodical and not unlike what the RAND people cooked up in the 1960s. They make investments in Africa and Latin America that benefit China, build up a military power as deterrence, and have a trade relationship with the US and Europe that transformed the nation from a decrepit backwater into a global power in 50 years.

Why would they want a pyrrhic victory (or defeat that would damage the Beijing government) in Taiwan? Taiwan isn’t Hong Kong. Aggressive rhetoric and policy accomplishes political objectives, and the long game is to foment instability in Taiwan and wait it out.

The risk of conflict with China is North Korea and Japan, not Taiwan.

Genuine question: does the elevation of Xi as (effective) dictator for life change the approach of the Chinese State to be less rational?

I want a graph with the dictatorshipness on X, age of the head on Y and rationality on Z.

That would require their government to have acted rationally in the first place. The first 30+ years of economic reforms weren't done out of rationality, but one of necessity to survive as the country was on the verge of total economic collapse by the end of the Cultural Revolution. Their hands were forced in the process of keeping everyone employed and fed, which gave a sense of rationality. Even during this supposed golden era of economic liberalism, there was rampant government corruption, oppression of dissent, and censorship (that was easily permeable because the technologies and funding for total blanket surveillance weren't ready until the mid-2010s).

The “this time it’s different” arguments also have a history of falling flat on their face. That was the common argument in Germany in favor of energy dependence on Russia, for example. It’s a bit wild to claim this theory is still legitimate when we’re literally in the middle of a shooting war proving it wrong.

It’s not like everyone just accepted the German policy as a smart idea.

They made a seemingly knee-jerk policy decision after the Japanese nuclear disaster.

The cause of WWI, if you can boil it down to a simple statement, is that red lines, vaguely drawn, were crossed by both sides, resulting in escalating tit-for-tats, with no one on any side acting to deescalate the situation.

The risk of this happening in a US-China conflict is moderately high. Chinese red lines are vaguely drawn. US red lines are more firmly drawn, but the US government system has multiple loci of power, and China seems to have a poor read on US policy as a result. De-escalation on both sides looks rather difficult to pull off, especially because China has a tendency to do things like through hissy fits whenever it feels slighted and shut down existing mechanisms to help de-escalate.

Another way a war can break it is something like WW2 or the Russian invasion of Ukraine, where you declare a war because you think you have a window of opportunity for victory. And if you guess wrong, well, you can ask Nazi Germany how well that works out for them.

That's because it's not a widely accepted term by any measure. I am all for Taiwan and its liberties, however, I cringe whenever I see journalists have made up a new term like "silicon shield" and expect people to adopt it.

This article [0] submitted a few days ago makes me think there are a lot more reasons than just chipmaking for defending Taiwan.

[0] https://news.ycombinator.com/item?id=34417336

> "The likely aggressors"

Is there any other than China?

If there are US fabs, and China invades Taiwan then Taiwanese fabs are destroyed, and China loses access to advanced chips, but the US and the west do not. This puts China at a major disadvantage compared to the US, for at least a few decades in the future, if it decides to invade Taiwan.

It seems unlikely that US and west would get involved in a direct major military conflict with China, no matter what. The costs are too great. It is far more likely that they would only provide military equipment and intelligence (like Ukraine).

So this move of TSMC actually strengthens the shield a lot more than it weakens the spear.

I suspect you are mistaken about the USA's willingness to go to war over Taiwan.

At this point, an attack on Taiwan is an attack on the USA. I suspect China is very much aware of this.

I suspect if the USA goes to war with China over Taiwan it will lose.

The USA would lose in Taiwan, or would lose a greater war.

The first seems reasonable enough given the proximity to China and Chinese hypersonic missles. The second seems rather fuzzy.

Doubtful because an attack on Taiwan would also be an attack on Japanese and Korean trade.

The three gorges dam is also hangs a bit of a sword of Damocles for china. A few good hits is as good as a nuke.

> At this point, an attack on Taiwan is an attack on the USA.

Who or what are you citing? That isn't US policy as far as I know. I can't recall any US president, political party, etc. taking that position.

I mostly agree with your points but want to point out there's always potential for things to change. The PRC is still investing heavily into developing native chipmaking capability in all stages of the production chain despite a string of high profile failures in the last 5 years. For military purposes the chips don't have to be made with state of the art technology. They just need to be good enough to work and achieve their intended purpose, sometimes rather crudely. Should a breakthrough happen, Taiwan's "silicon shield" could weaken quite a bit.

I still lean towards a "no hot war" scenario because the PRC's demographic structure is deteriorating just like Russia's, and Putin has already demonstrated a lesson with the war in Ukraine.

In practice, if war happens and TSMC facilities are bombed to avoid passing into Chinese control, how many years does the world go back in processor state of art? Are we back to 486 level, Athlon 64 level, core family first generation, or what? What is the new average feature size, in nm, we are likely to fall back to if things go south in Taiwan?

It is just a few years on node size (Samsung is close), but probably large setback on production capacity (Samsung has less volume).

When it comes to international security, I don't think those in charge think about TSMC more than just in passage. Just look at Ukraine. The US is helping Ukraine a lot, but there's pretty much nothing that Ukraine is/was producing that the US needs.