TSMC says efforts to rebuild US semiconductor industry are doomed to fail
appleinsider.com
appleinsider.com
Look at Ukraine, aside from grain and a few specialty items, it's not a huge exporter to the world. Yet a serious war there has caused global problems. China going for an expansionist romp into Taiwan will be 20X as impactful.
Anyway, I thought things were going to go a lot different in 1999, last 20 years have been a real ride.
So, the initial invasion might just fail, replacement material will come over the Pacific and from Japan, global trade will basically cease to exist. On top of that you have the fallout from a military failure in China itself. In short, it'll be a mess of epic proportions and Taiwan will most likely not be invaded successfully.
And ultimately TW doesn't have to be invaded successfully, it can be blockaded permenantly (and trivially) to deny both war material from going in or chips from going out.
@hef19898
> potential to start a shooting war
Escalating to shooting war is expected, but the point is a comprehensive blockade by PRC is feasible worst case contingency plan that can be done without any complicated hardware. Glide naval mines and MLRS can be launched from mainland in quantities that US+co simply lack capacity to circumvent to thoroughly lockdown TW ports/coasts and airstrips indefinitely.
@LAC-Tech
>yet don't hear a peep about Chinese military issues on Global Times or XinHua
Listen harder? State media like Xinhua or CCTV7 (military) shows PRC accidents all the time. CCTV7 interviews also frequently shit talks about PRC military deficiencies, but it's dressed up as we've strived hard to improved substandard X to modern standards etc, and we still need to improve on areas we're lacking.
How much did you hear about Russian military issues in January?
It can be seen as the inflection point where China realised that churning out vast numbers of obsolete Soviet designs had to yield to quality and training.
This was Russia's exact hope in Ukraine.
China needs to do about a millions things very competently and quickly to take Taiwan. The activities involved are things that would be hard for most western militaries today, even with all their money and skill.
Here's a list of difficult things for china to do: 1. It's got to stage it's troops in a way that's real but maintains the element of surprise against an actor with US Intel backing it up. Even if it's just 1 guy from the NRO with a satellite phone and a laptop. The edge for Taiwan will be massive.
2. It's got to be able to cross 100km of open ocean with about 10-20 thousand troops as a first wave that can be dropped off rapidly to fight and survive on it's own.
3. Then those same ships will need to return and onboard more and more waves of troops that will need to be landed until they can safely establish maritime operations. In land warfare gaining air superiority is a first major goal. This is one of the reasons the Russians have gone so badly in Ukraine. For China, they will need to be able to own the sea and provide safety for their own ships as a highest priority.
4. The fighting on the beaches, will turn into fighting in the cities, and then fighting in the hills. Each one of these is a real challenge on it's own. It took the US military years to get good in Iraq.
5. Whilst China has lots of cash, and lots of men, and lots of tanks etc. The will power to fight may be an issue. For the Taiwanese soldiers, fighting and dying may mean independence for their family and nation. For the Chinese soldiers, fighting and dying probably just means they die. Its kind of a pointless activity for the soldier on the ground who just needed a job when they signed up for the PLA.
Why would China use tactics and weaponry from 80 years ago though?
China has lots of medium-range missiles, but that won't help much at the beach.
Bullets and artillery are much the same. But AI weapons such as anti-ship missiles and the slimmer 20 inch diameter torpedoes are more use on the defense.
It would be the same for any operation against Taiwan. None of the steps you've outlined would happen before achieving air supremacy.
Taiwan has a weak submarine force, so the Chinese objective would be to control the air over the Straits to prevent the surface fleet intervening.
Once air supremacy is achieved in any conflict the dominoes of layered defence start falling.
I would imagine that any Chinese move towards Taiwan would be preceded by a couple of weeks of ferocious air attrition and ballistic missiles being used for SEAD.
It turns out the 90s was basically western civilization's peak.
This is a surface level take on the situation.
The disruptions to global shipping would be enormous, to say nothing of security implications, or fear of open conflict between nuclear powers. Potential destruction of semi-conductor fabs would be just the tip of the iceberg.
That's pretty rude and meaningless. Why not just say "I disagree"?
> The disruptions to global shipping would be enormous, to say nothing of security implications, or fear of open conflict between nuclear powers. Potential destruction of semi-conductor fabs would be just the tip of the iceberg.
Not if the rest of the world does not come to Taiwan's aid and allows China to roll over them, obviously. That is presumably what OP is saying their worry is, so that is what you have to address if you want to disagree.
"Surface level" invites someone dig deeper. It's not a personal comment about a persons latent talent. If you think it's rude to point out a common opinion based on surface level knowledge of a complex situation, that's on you.
We have the ARM licensing example. Could a player like TSMC do a reverse license where they license out their FABs factories for different other players? Or is that at the core of what they do but in different form?
I think TSMC knows that tech is at the top of a pyramid that relies on industrialization and military might to secure it. IMHO we are going into a very unsecure future where we will see an accelerated decline in all the factors that contributed to globalization. In particular, I worry a lot about food security and supply when global markets in fertilizer are disrupted, energy security when global markets in fossil fuels are disrupted, and manufacturing of all goods when commodity markets are disrupted. As much as I want a clean energy and EV revolution in my estimation, it cannot happen without Russian raw materials.
Furthermore, unless there are amazing leaps or breakthroughs in energy generation, energy storage, and food production without fertilizer, the rest of the world—but not the US—will be in for an uncomfortable few decades.
TSMC executives will of course try to influence US politicians, but I think US politicians know that US voters are fatigued with foreign military excursions. US active-duty numbers are already historically low, and I suspect this has to do with difficulty with recruitment. I think it is fair to ask, "Am I defending my country by fighting China in Taiwan?"
TSMC’s assertion “We’re the best, don’t even try competing” strikes me as “fear masked by false bravado” - a bit pointless, really.
Saudi Arabia’s king can give their people all the money they want, but they will not be able to buy the technical and management expertise and culture that takes decades to develop without waiting decades for it to develop, with the risk that it does not.
Sufficient is an important word, and sufficient expertise is not abundant in the business this thread is referring to. That is why Apple can lock up the most advanced chip supplies, that is why only one business in Taiwan makes it, why only one business in Netherlands can help the one in Taiwan make it, and why only one business in Germany can help the one in Netherlands make it.
Not that Apple and USG are not more or less indispensable than TSMC/ASML/Zeiss, that is neither here nor there. The point is some things simply have no fungibility, and hence you cannot buy your way out of it.
If 2 billionaires got trapped on an island with all their wealth in cash, and neither knew how to build a boat or sail, all that cash is not going to help them avoid having to figure out how to build a boat and navigate the seas in sufficient time to survive.
My argument is it is possible the US cannot throw money at it until it is resolved, perhaps due to the R&D not resulting in success, due something external that causes US to lose ability to throw money at it, due to something internal that causes US to not politically support it, enough people not sufficiently interested in the sciences, etc.
They were so appauled at the political disfunction, they quit the project. They started one in morocoo instead - its already operating while the californian one is nor even 50% complete.
Your political disfunction is part of the package. Incompetent management is too. You cant just hire 1 expert software engineer into dysfunctional conoany and expect great results
Culture matters more than money and Western culture is on a downward spiral.
Culture does matter, but its pretty clear the west is the entrepreneurial engine of the global economy and that doesn’t seem likely to change.
At the same time Portugal was a bunch of poor fishermen trying to get past a sandbank on the African coast.
If US has destroyed its human capital in chipmaking, it will take decades to rebuild.
And while doing so, it must navigate the same political system that makes it impossible for California to build a train
Money and fear, as we have seen, is a powerful motivator.
They are always trying to poach engineers from TSMC,Samsung and others by offering N-times the salaries and it resulted in no significant gains.
You can't just throw money at people and get them to come work for you wherever you are. Only some people will take the opportunity, others will decide other things are more important than money.
The Russians also assumed that "throwing money" at problems, like maintaining a huge army, would somehow resolve them without having to actually do the nasty job of, you know, solving problems at their roots.
When push came to shove, it turned out that all the money went to some place nobody knows, but the problem stayed and got even bigger.
I cannot even get my mom’s baking/frying recipes to produce the same taste and textures that she can, yet I was the one watching and documenting them multiple times.
The US is not capable of having a national policy on anything, because power is too fractured and fragile. Which is not a bad thing overall necessarily, but it is a bad thing in this case.
Taiwan is a small unitary state, as is South Korea. The governments of both know that the national interest and their chipbuilders’ interests are aligned. In Taiwan’s case especially so.
There is always a huge gulf between what modern nations COULD do and what they actually WILL do. In the west, we're allergic to making any sort of painful decision or any sort of delayed gratification. That's why the war on terror has blown back so hard. Could the Nato liberate Afghanistan? Of course! Did it? No, we spent 20 years mismanaging it then gave up.
Personally, I'd put this in the same category as Universal healthcare or green energy or fixing education. Of course we can do it. But we won't.
>Backed by the might of its own economy, surely the US can rebuild dominance in any technological area it happens to choose given its strategic position and aims?
This is why I don’t worry too much about Chinese saber rattling towards Taiwan. Imagine the Chinese economy with 92% of the world’s supply of advanced computer chips gone.
People like Erdogan and MBS, sure, those are just irrational, deluded bullies. Apparently Putin and, it seems, Ping, are the same
As opposed to British housing market, which has been called a housing crisis for the past 10 years?
And our handling of covid where hundreds of thousands of people died needlessly, and elementary steps like improving ventilation or air filtration in school and large public buildings are not taken.
From Brexit, to Afghanistan withdrawal, to energy dependancy on political opponent, I am under impression that decision making of western nations is more chaotic and erratic decisions than China's.
We seem to take it as a given that as political leaders change, our decisions are erratic. But when China's decisions are erratic, that's pointed to as surprising or as failure. But Xi is not personally in charge of everything, there are various leaders and influencers withing the CCP and they change.
Ah yes, good old American Exceptionalism. Get yourself into another war that you need to get half the western nations to help you get out of again, why don't you?
America is the nation stopping this happening.
When the US goes to flatten countries in Asia, the Asian powers politely ignored them. There were some diplomatic protests. The war stayed regional. There is an argument that this is the better approach.
There is room here to complain about the US even if they are the main defender of Taiwan. It is good that they are promoting prosperity in Taiwan, but their habit of escalating conflicts could literally get us all killed.
As to your specific charges "When the US goes to flatten countries in Asia, the Asian powers politely ignored them.", you'll have to be more specific.
Iraq, Afghanistan. Threats against Iran too.
And if it is anything like their approach in Europe, they'd probably move in and escalate. So in a scenario where China invades another country and triggers a global economic meltdown it is still possible that the US would swoop in and cause the most damage. The Ukraine thing would probably be all over by now if not for the US deciding to set up a proxy war on Russia's borders. I don't think there is any contention on that either, the only argument is whether it was a good idea or not.
If the US wasn't involved in Europe, it is likely that there wouldn't be a crisis of the magnitude that is unfolding. When the US invaded numerous middle eastern countries, the Asians pursued policies of de-escalation (realistically, there were no alternatives) and the the crisis stayed relatively contained.
That's like saying the Iraq War ended in 2003.
The death and repression would just be getting started.
No western nation wants China to forcibly take TSMC. None
On the other hand, the PRC needs to build their own semi industry, which they are doing, and which American sanctions are pushing them to do, up to the point where the potential cost of losing Taiwanese supplies is 'acceptable'.
Ultimately the aim of re-unifying the whole of China is a deep national and political undertaking, not an economic or financial one, so a level of economic/financial cost would be accepted.
Westerners are incredibly selectively blind when it comes to nationalism. They pretend no one believes in it any more then every 20 years or so they blow up the world with it. Last time was 9/11, before that was German unification.
I can't fathom the self delusion this takes.
In 1989? Or are you making a cryptic reference to 1871...
I know this is someones "I'm too scared to be edgy on my main account" account, but still, make it clear what you are referring to!
PRC economy will suffer much less than US economy with 92% of the world’s supply of advanced computer chips gone. Considering how US reaps significant value add from IP through the entire semi chain. Or relative loss for US high tech with greater market caps. Or that PRC strategic industries are losing access to those chips due to US export controls. Pretty soon PRC dependency on TSMC will be reduced to point where it will be in PRC interest to disrupt TW semi because world where US has access to 8% leading edge semi and PRC 0% is preferable to one where US has access to 100% but PRC... still somewhere around 0%.
The semi industry in Taiwan and Asia in general has been a boon for the world, with fast, high-quality manufacturing at great prices for anyone who needs it.
The US has always sought to influence this to some extent with sanctioning countries, and threatening the asian semi industry directly with further sanctions if they were to continue to sell to countries the US don't approve of.
This also highlights exactly how the US views and uses its own semi industry: as a weapon and means of control by dependence.
A strong asian semi industry would mean slightly higher prices for US American consumers, but it would certainly be better for the world as a whole, and that's what really matters.
Any business anywhere: "Oh, our future competition? Yeah, you should not go with them, they are doomed to fail".
That is standard operating procedure.
>> When US House Speaker Nancy Pelosi visited Taiwan in August, she met with Morris Chang and Mark Liu, chair of TSMC. Chang told Pelosi that Washington's efforts to rebuild its chip manufacturing were doomed to fail,
What's the fundamental difference between building a fab in Taiwan, US, Europe, Israel or wherever else besides for the ongoing labor cost?
2. It's also significant that Morris Chang openly is saying this. Publically, and repeatedly. And possibly specifically to Pelosi given the politics behind him merely being in the same room as her.
Important read between the lines consideration is whether Chang's attitude will affect success of TSMC fabs in US. Apart from being burnt in past, he has also been consistently against building leading edge fabs abroad and IMO maybe personally invested in undermining unprofitable expansion efforts (without even going into geopolitics of silicon shield). The consensus when rumor of TSMC building fabs in US (pre CHIPs) was that it made no sense because TSMC already announced record CAPEX investment in TW months prior, and there was no money for new expansion let alone leading edge. Post confirmation of Arizona fab and so far the sentiment I've gleaned is that the CHIPs subsidies to TSMC + culture differences still makes the expansion not worthwhile for Chang.
That's just the practical operations/economical considerations. Now layer on the geopolitics, i.e. eroding TW silicon shield, behind scene US pressure to get these fabs, unilateral export controls disrupting/severing TSMC's PRC business/interests. TLDR is maybe Chang is just really (understandably) annoyed at US fucking around with his company and maybe even him on an individual level considering there was probably some sticks being whipped behind the scenes especially if the CHIPs subsidy carrot wasn't enough.
Question is if that will affect success of TSMC fabs in US. I'm sure you'll find similar feels in board rooms of Samsung/JP equipment makers. There's a reason US had to announce new unilateral export controls after talking up / pushing through CHIP4s as collective effort to constrain PRC tech, Asian semi countries have no interest in US gaining semi independence/dominance at their cost.
E: To add, US discussion around TSMC has gotten to the absurd point where TW media in the past months had to retort: No US Army War College, we're not rigging fabs to explode; and we're not goin gto cooperate with paperclipping all our semi engineers in event of war. We're going prioritize evacuating grown ass men over women and children first like any other sane planners.
TSMC is a strategic asset and most of its value and leverage it confers stem from it being physically located in Taiwan. That's exactly why the US are pushing for them to build more fabs in the US and why Taiwanese authorities are most probably not keen but they also need to preserve their relation with the US...
But I'm confused, because the title is: "TSMC says efforts to rebuild US semiconductor industry are doomed to fail"
It's obvious why they want to say that but nowhere in the short piece is there a single word as to "why?". Doom is not discussed. Nor in the linked Taiwan Financial Times. Morris Chang must have some sort of theory, right? Economic? Technical? Political?
Maybe its that TSMC just manufacturer stuff. They transparently produce silicon for anyone who wants it around the world.
Intel (read the many recent HN threads here about management engines, non-transparency etc) bake political control agendas into their silicon.
Biden should spend that $50B on creating a portfolio of smaller, independent rivals to AMD and Intel's fabrication. With state of the art fab plants that just make what their customers ask them to we can let the US/Euro market in silicon IP flourish... RISC-V and much more.
Propping up behemoths that have wonky agendas, lines of old and failing products, and refuse to listen to what markets really want might be the "doom" Mr Chang is alluding to.
The US's wage levels and infrastructure costs are too high to be globally competitive enough for success, apart from the over-valuing of the US Dollar.
Tl;dr
Doom is implied. Never spoken of. Generic story refactored you’ve read a million times about Morris founding TSMC. Geopolitics implied. TSMC says it’s doomed, while it helps the US make plants.
I am not worried about China invading. It’ll be over for them. Even if they take over Taiwan with no casualties on either side and everyone wanting to make chips as usual, they still fail without Dutch lithography which is like TSMC for another required process. https://thediplomat.com/2022/08/lithography-the-achilles-hee... They don’t have the high quality silicon required either, that comes from the US.
You will be downvoted, flagged, and then dang will arrive to warn you that this specific argument is not permitted on HN because it inevitably leads to unproductive flamewars.
To be fair, dang is completely correct in this observation. It never leads to productive discussion, but it is nonetheless verboten.
Regarding the cultural reason: Where is the obstacle? Somebody mentioned that TSMC receives the wavers from the USA. Why would it be possible to operate one set of clean-room machines, but not another? If I am not mistaken then there are also the Intel plants. They may have lost their lead, but they are still operating the machines.
Either Taiwan gets a nuclear shield which is NATO alike (a country ready to die with them in a nuclear war) or it is a "done deal".
I guess leading-edge chip manufacturing will have to move in a country without an authoritative regime and with an explicit nuclear shield.
From Chinese Anti-Secession Law:
Article 8: In the event that the "Taiwan independence" secessionist forces should act under any name or by any means to cause the fact of Taiwan's secession from China, or that major incidents entailing Taiwan's secession from China should occur, or that possibilities for a peaceful reunification should be completely exhausted, the state shall employ non-peaceful means and other necessary measures to protect China's sovereignty and territorial integrity.
https://www.europarl.europa.eu/meetdocs/2004_2009/documents/...