If countries are less reliant on manufacturing within Taiwan, its one less reason to stand up to Beijing's one China principle.
Yes, they depend on ASML, but ASML itself was dependent on IP from TSMC engineers. TSMC already announced that they will stop applying for subsidies from the Chip Act[] in the US recently because the US keeps changing what confidential data needs to be shared with the US.
They seem to be at least slowly understanding that this is a dangerous game.
And for the people that keep talking about ASML you might well want to remember that not long ago the leadership was mocking China's capability to build their own domenstic supply chain only to flipflop shortly after saying that it would be foolish to abandon the Chinese market likely indicating that they might be concerned that their domestic supply chain might end show up faster than expected.
Taiwanese media last year was plastered with news about how the US hollowed out the Japanese semiconductor industry with its agreement in 1986 and how that will be potentially the fate of Taiwan.
But the sibling comments are correct about the potential blowback, let's not forget that the US has been talking about bombing TSMC themselves[]. I guess that probably explains why they just announced a delay to the construction of the Arizona fab.
[] https://www.koreatechtoday.com/south-koreas-semiconductor-ex...
[] https://www.aljazeera.com/economy/2023/7/27/samsungs-profit-...
Well, ASML is primarily and rightly very concerned that once "China" has their machines they'll attempt to reverse engineer them, ignore any western patents and IP and try to build 'Chinese chip machines'. It's happened to many industries, from Lego to fashion to hi tech. Hell, apparently there's even Chinese Knock-off Movies.
So, above everything else, there's a legitimate concern for ASML that once they move or deliver too much into China, they'll enable their own competition.
There's over a 100K parts, several of them from exclusive suppliers. You'd have to recreate several industries from scratch or somehow bribe all suppliers. You will absolutely fail to recreate the parts at all but even then if you hypothetically would, you can't put the machine together as if it's just a few bolts. It requires a team in the know months to do it, but you're not in the know. The tolerance for error is near-zero. Installing, configuring, running the machine, both hardware and software is extraordinarily complex.
None of this is a secret. The Chinese government announced a multi-billion dollar program to try and recreate such a machine from scratch. Expected timeline is 20 years with a highly uncertain outcome.
ASML does not have a concern to export to China, they want to export to China but are pressured to not do so by the US government.
Probably not as carbon-copy. But certainly important parts, or older versions.
ASML isn't just selling a millions-dollar-machine, they sell maintenance, support, upgrades, refurbishing. All these pieces are important in themselves, not just the whole assembled machine. Any single piece that can be reversed-engineered can be a danger to their moat.
This isn't just ASML, this is any hi-tech industry. I worked in power-plants, where ABB didn't just sell us gas-turbines, they sold a package, from Swiss engineers coming twice a year to help us tune them to "we now have these nozzles that make them 1.2% more efficient in hot weather, we can come replace them". Or the story on the HN frontpage recently, where Russian airlines are running out of brakes: just imagine being, say, a Russian factory that makes, say, brakes for Tupolevs and you can reverse-engineer Airbus brakes and start making them. High margins guaranteed.
So yes: ASML is rightfully very careful with Chinese competion and espionage¹
¹ https://nos-nl.translate.goog/artikel/2280228-wat-heeft-de-c...
Dont assist even your incompetent adversary. Competition 101.
The U.S. said nothing of the sort. One Congressman, in the minority no less, did.
But an idiot fringe congressman spouting such things does not make a credible threat or say anything about the actual intentions about the country, just a good way for somebody to get attention or some points from one hardline group or another.
Is demand for electronics finally dropping? I want to be able to buy a 3 year old Nvidia low-range GPU at below MSRP...
https://www.extremetech.com/gaming/report-nvidia-has-practic...
So I wouldn't hold my breath.
Unfortunately, it is likely to be limited to flash storage. The $/GB value has fallen to historic lows for all manner of solid-state storage devices.
> "“There is an insufficient amount of skilled workers” with the expertise to build a chip factory, TSMC chairman Mark Liu complained during a call with analysts. The executive warned the company might have to fly in “experienced technicians from Taiwan to train the local skilled workers for a short period of time.”"
https://fortune.com/2023/07/21/tsmc-complains-cant-find-enou...
The neighborhood around ASU in Tempe, AZ is chock full of fabs that have been built or expanded recently. Please take the “hollowed out workforce” to a different discussion where it actually applies.
That's a very strong prerequisite though… It may not be over for several decades to come.
- Hong Kong belonged to China already, it just had a special status that was supposed to expire at some point, Xi “just” went ahead of schedule but there wasn't anything to stop him from doing so (no independent HK government, no army, etc)
- Taiwan is a complete state, with a government and an army, a navy and an air force equipped with Western and mordern home-grown equipment, cruise missiles and anti-ship missiles and so on. Without a Western direct armed intervention, Taiwan is almost always expected to fall after a few weeks by Western military analysts (though the same analysts also expected Ukraine to be unable to match Russia on conventional battles), but even the current predictions on Taiwan suggests that China will pay a signifiy cost in terms of casualties and and material even if they succeeded eventually. And again the success assumes no military actions from the West. Of course Xi may take his chance anyway, but the situation would still be incomparable to HK.
Hong Kong was leased to the British Crown in the perpetuity.
The new territories had a 99 year lease and in 1997 had more than 2 million Hong Kongers living there. About a third of the population.
The more appropriate term is probably foreign direct investment (FDI) in Germany from TSMC.
I find it funny that western people have issue with the idea that Asian companies can indeed also invest in western countries. It is akin to some discussion about Lega here on HN where some people talked about it as "onshoring" when they made a US factory – Lego is a Danish comapny.
TSMC alone has no control over if the Chinese government will commit such aggression, if they would do their whole company strategy based on fearing the CCP, there's no way they can be successful.
They are willing to burn a lot of money and time to reach their overall political goal of annexing Taiwan even if it takes a very very long time.
Ultimately it's unrealistic to assume that a country of 1.4 Billion people can not catch up eventually espescially if they would otherwise have to rely on a global competitor (US) or a military target which factories are rigged to blow (Taiwan).
The best thing Taiwan can do (once their semiconductor lead is no more) is to make themselves into a strategic ally located on Chinas doorstep.
East Germany famously tried really hard to develop a native capability, but failed and wound up smuggling parts from the west and running years behind western firms: https://www.youtube.com/watch?v=cxrkC-pMH_s
China is more willing to both reward and hold senior stakeholders responsible, as long as they aren't too closely allied with Xi
Why? By 2100, each of Pakistan, the D.R.C. and Kenya are forecast to have larger populations than America [1]. Is it unrealistic to assume they won’t be at parity with the west in eighty years? (Note: not saying dismiss entirely. Just baselines.)
China has a lot going for it in semiconductors. Demography isn’t one of them. The population is aging and shrinking. That, ceteris paribus, reduces surplus capital for R&D.
[1] https://en.wikipedia.org/wiki/Projections_of_population_grow...
Pakistan has the best chances to rise, at least assuming religious / ethnic tensions can be somewhat placated, it doesn't become a hotbed of natural disasters like last year's floods and they don't either get into open war with India or suffer from some sort of coups by the military or secret services.
Kenya and especially Congo? That's a different game entirely. Climate change, utter poverty and endemic corruption are just a couple of the hurdles to pass.
Because PRC is not systemically comparable to those countries just like Japan isn't comparable to a larger Nigeria.
Only the most cognitively challenged demographic determinists argue larger population = more capability. Useful demographic analysis more than naive reading of demographic pyramid, it's recognizing unlike most developing countries, PRC has established record to upskill/cultivate and coordinate human capita at scale. And is projected to have multiple times more skilled talent than US by 2050s, moving from current 25% high skilled workforce to 60/70/80% to reflect workforce composition of advanced economies over time.
>Demography isn’t one of them
More than anyone, demography is PRC's greatest advantage for semi and other high tech industries in relevant timelines we're talking about. The workforce for high skilled talent is exploding for another 30+ years. Short/medium term PRC has by relevant measures, the greatest skilled demographic divident ever, and even in stagnation they'll be settling with largest (or second largest relative to India) pool of skilled workforce, with again proven ability to coordinate (i.e. not India performance).
Even 2100 PRC working age population is still ~2x of 2100 US if you tally up pyramid projections of 20-65 yr olds. Except PRC's 2100 workforce will likely have more than 2x current skilled talent instead of being dragged down by 100s of millions of farmers and surplus informal workers who will age out post 2050s to be replaced by better educated/more productive cohorts. Same way other Asian tigers grew despite "bad demographics" - by continuously upskilling % of workforce initially dominated by peasants over time.
This is why current analysis of US semi workforce project 80k short term shortage to balloon to 300-400k medium term. SKR, TW both have current/short term 50-100k shortage. The TLDR is SKR, TW (and JP) have nearly satuated their workforce in terms of skilled talent, and can't replace at parity. US expanding fabs as rate their talent pipeline can't catch up, and the ability to import/brain drain is limited because they'll be taking from other CHIP4 partners also undergoing shortage. Which leaves PRC, who after elevating semi to first-level dicipline in 2018 is pumping out about 30k IC graduates per year. They're still about 200k short, ~520k/720k (400k in 2018) out of what IC talent 2018 white paper estimated PRC needed for completely indigenous semi industry, which US isn't currently even trying to pursue, still ultimately relying on east asian semi supply chains. Long term US may to get there, but right now only PRC is sent to close semi labour force gap within the decade. Ultimately, talent is just as important in industrial policy as money. PRC has both. Their biggest shortcoming is being behind, which will likely be addressed by investing sufficient talent and money.
https://en.m.wikipedia.org/wiki/List_of_countries_by_governm...
The Taiwanese government can attempt to prevent TSMC in participating, but what good will that do? Might as well ensure that TSMC at least have a hand in it and avoid customers leaving for another manufacturer. TSMC can then offer fabrication, maybe it's in the EU, maybe it's in Taiwan, the important part is keeping the customer and not allowing competitors to establish themselves.
This in Germany and the one in the US are both Megafabs.They produce chips for special industries with strategic importance. In the US for military and defense, in the Europe for automotive industry and automation.
Having this beefy reliance on a small island that gives china the middle finger
-OR-
derisk the reliance on taiwan
Stopping cargo going out of Taiwan is rather cheap. Unguided rockets and aircraft from your homeland can do a lot. And then they just have to defend near their home land while waiting for their opponent to run out of resources.
At the same time, as we can see here, it takes 4 years to go from acre to functional fab plant after planning and negotiations. And any kind of war to regain control of the chip supply lane is a significant drain on the chips and systems reliant on the chips you have.
Millions of lives would be lost if not more if it will be a full scale war. A chip fab is nowhere the same level of importance to be even discussed on the table.
"But they've spent X billions of dollars." Yes but this is irrelevant when facing off an existential threat to their business and the overall economy and security of Taiwan. Taiwan, like Israel, is run by cunning folks. They wouldn't have survived this long otherwise.
Mere reduction of the criticality of Taiwan's advanced manufacturing will not eliminate Taiwan's geostrategic importance.
Plus, the CCP's insistence on being an expansionist authoritarian state is reason enough to contain that expansion, to prevent further resource gains.
Sure, if you ignore their military actions and treaty violations in Tibet, Turkistan (Uyghurs), border conflicts around the Himalyas and India, Russia (border conflict), Korea, Vietnam, everywhere on the "9-dash line", Canada and the US (where they are putting up remote "police stations" to intimidate anyone expressing opinions they dislike)...
CCP has a highly consistent history of being as aggressive and untrustworthy as they can get away with. Anyone trusting them is a fool.
ASML definitely wanted to sell to China to increase their quarterly revenues but couldn't because the US owns the IP for the EUV tech and ASML's EUV light sources are made by CYMER in the US, which ASML bought but they're still under US trade restrictions.
So no, ASML and the NL government have to dance according to the moves dictated by the US gov as they own the keys to EUV tech, not ASML/NL.
But your details aren't exactly right, I think. Cymer builds light sources for DUV, not EUV. The EUV light sources are built by Trumpf in Germany, at least as far as I know.
https://www.asml.com/en/news/stories/2022/making-euv-lab-to-...
> at least as far as I know.
You could have also googled this to factcheck before posting if you were unsure ;)
[1]: http://www.transientek.com/cpzx/jcxnmgtjgq/jzwckgy/35.html
[2]:https://thechinaproject.com/2023/08/03/beijing-pushes-china-...
But as we saw with the Evo Morales grounding incident https://en.wikipedia.org/wiki/Evo_Morales_grounding_incident there's still a certain amount of jumping.
Also remember that a lot (all?) of the circuit design software is US-based, so the supply chain could be squeezed from that direction as well.
By "exports", do you mean donations? https://www.bbc.com/news/world-us-canada-64656301
For example, last year the US sold $6 billion-worth of Abrahams tanks to Poland. Overall, paid-for exports to Europe have exploded.
When you said China, which China are we talking about?. the fact is, Taiwan and China is still in a Civil War.
The United States did not explicitly state the sovereign status of Taiwan in the three US-PRC Joint Communiqués of 1972, 1979, and 1982.
The United States "acknowledged" the "One China" position of both sides of the Taiwan Strait.
U.S. policy has not recognized the PRC's sovereignty over Taiwan;
U.S. policy has not recognized Taiwan as a sovereign country; and
U.S. policy has considered Taiwan's status as unsettled.
The PRC has never, not even for a single day, controlled Taiwan, Kinmen, Matsu, etc.
The continued existence of the ROC (1912-present) truly causes tankies to lose their minds.
Anyone can talk a big game.