One thing that was not mentioned was MAO 3.0's (a.k.a., Xi Jingpin) ambitions to realise China as, again, the "Middle Kingdom". That is to say, as history should have taught us, a dictator's bent should never be discounted -- irrespective of the costs incurred to achieve it.
No country can survive without TSMC. If China did take over Taiwan, the world would have to make peace with them, because it absolutely must have those chips. And by the way, the takeover might be more like Crimea than like Poland in WWII. China just takes over without a shot fired after a campaign of increasing pressure and assuming the result. Alternatively, they might have to take out Taiwan's air defenses/air power first. Then they can flood the zone. How much damage will the Taiwanese be willing to accept to avoid their fate? (No idea.)
Thanks for the wonderful piece. I think pushing the economics a step further is important: if TSMC suffers via a loss of intermediates or a loss of customers, the company shrivels. If that happens, the staff become poachable, and throwing money at them helps move the expertise to the USA. You might need Walter White to build the fab, but if TSMC shrivels, Walter White becomes available.
All the processing equipment in the fabs comes from the US, Europe, or Japan. Each equipment manufacturer keeps a small contingent of home country engineers overseas, on rotation, to do any serious maintenance on the equipment. Without those overseas engineers, the overseas equipment knowledge, and overseas parts, none of the processing equipment continues to run. When the equiment cant be maintained, the fab cant produce product in the long term. My guess is the long term is on the order of 100 days.
Also, like there is no shortcut to go to the next node, there is no shortcut to build the next generation of processing equipment. The process machine knowledge has built up over generations of engineers and has been refined through feedback at each node of in-fab operation.
Semiconductors are just an example of the general point that it takes a world of specialized knowledge and inputs to produce any high technology product, as you suggested with your Xilinx example. Like it or not, no country is going to be able to have both high technology and independence. We are all going to have to get along, or go without.
China does not need to kinetically invade Taiwan to bring it to heel.
It can simply assert its UN-recognized rights to dominion over the island (acknowledged in Taiwan's constitution), close the Strait, and assume control of all air and sea traffic into and around the island province.
Unless the US is willing to lose several West Coast cities in a hurry, there is nothing it can do to stop China, whose military is far more powerful in the region. Nor is are Taiwanese eager to sacrifice their lives defending a corrupt backwater when they can make better money under the PRC.
After that, it's a matter of bargaining, and the Chinese are pretty good at that.
Nice article. One thing though. There is only one unique critical supplier: Dutch ASML for EUV. For all other equipment there are several suppliers. Applied Materials is by no means unique. ASMI And Be equipment are just as good. The US only has leverage by pressuring others.
A Chinese invasion of TW would destroy TSMC - the short term impact of the loss of that chip supply would be devastating for the world economy, but disproportionately felt by the Chinese who will have no alternatives other than SMIC (which will be cut off from SME from the West), while the US can still source from Samsung / Intel (and very likely will provide very heavy pressure on Samsung not to supply China). Medium-term the US is in a better position to stand up to new foundries in the US to produce those chips (Intel, possible Samsung foundries in the US), while it will take China much longer as they will have to replicate the entire semi value chain AND catch up the head start that TSMC/Samsung/Intel has over SMIC.
I also can't see how there is any realistic possibility that China either captures TSMC intact, with no damage, with no loss of IP and knowledge from critical employees and vendors. Or how China has even a shot at coercing TSMC to supply China and not the US (as you stated, all the SME inputs other critical technologies into TSMC are controlled by the US)
China's only route to semiconductor independence is the long route - the investing of hundreds of billions of dollars over a decade or more and reproduce every step of the chain while trying to catch up on the tech frontier. Which would argue against China initiating any attack on Taiwan except in the most desperate of circumstances.
Hello, sir, you may have overlooked a problem: Taiwan has been China's territory since ancient times, but Taiwan has been colonized by Japan for a period of time, resulting in China's reunification. So it is very inappropriate for you to use the word "invasion". Thank you.
I just want to know which US Semiconductors stocks to invest into in case of a Chinese invasion of Taïwan. When TMSC goes offline, US chipmakers will pickup the slack and their stock price will go up. Please tell us that. Where to invest our money ?
What about the option of the US destroying say 4 leading edge TSMC fabs in Taiwan, as soon as they "discover" the CCP are invading? Disturbing the leading edge development of Semis for say 2 years, but ensuring the CCP does not hold the reigns to TSMC, and others picking up the slack.
CCP could then focus on all the human and natural treasures that Taiwan has to offer them...
You don't think this is a realistic scenario?
So what if for 2 years we can't make better graphic chips? Have to make cars where the seatbelt doesn't beep? Life goes on. Growth slows- ALOT- ok. CCP however NOT running the tables on the semis front...
I would think it would be possible to get most of the TSMC workers out of Taiwan by plane or boat if China invades. With all the equipment being produced in the west they could set up shop in the west. Fabs can be built in advance so they can move into ready facilities.
Taiwan and China are already tightly coupled. Taiwan is represented in China's Congress, for example, and its freighters uniquely enjoys the right to sail on China's vast internal waterways–giving it huge time and cost advantages over rivals. President Xi has relatives in Taiwan and the Taiwanese trust him more than they trust their own politicians.
My Chinese friends laugh at the idea of a purely military takeover of Taiwan. More likely, they say, is a bland official announcement, "As of midnight tonight, all movement of goods and persons to and from Taiwan must clear PRC Customs and Immigration. Otherwise, no changes are contemplated".
Of course, this would be enforced by China's immensely powerful military and a picket line of missile patrol boats around the new ADIZ. The Taiwan Strait would, a fortiori, become domestic waters, closed to all military traffic.
In other words, business as usual, except that global power would shift decisively in China's favor both through its control of IC imports and exports, and the impotence of the US military to hinder the move.
Question: I get your point, and agree, that tools suppliers like ASM Litho, KLA- Tencor, Applied Materials and Lam Research are critical to any fabs ability to operate at the cutting edge. Just curious, where do the US based design software companies Synopsys and Cadence fit in on all this? Judging from their incredible margins and my understanding of the business it is basically impossible to design leading edge, or maybe any, chips without their software.
This article seems to be based on the dubious assumption that China wants to invade Taiwan. China wants to gain control over Taiwan in the long run and invading it is not the most convenient way to achieve such control. Constant military pressure with minor clashes is Chinese optimal strategy. If there will be war, then China will destroy large part of Taiwanese military in one day without invading Taiwan. The fact that Taiwan can not defend against China will become completely obvious in the coming years and China will slowly establish political control over Taiwan.
OK in parts, but some really unforgiveable gaffes. The author at one point makes the correct point TSMC can't function without continuous inputs from US companies and those that must follow US dictates, like Japanese suppliers of semicon raw materials. The US can shut down all production, yes all production, at TSMC at any point and the fabs in Taiwan would no longer be able to operate. Not a problem money can solve.
One thing that was not mentioned was MAO 3.0's (a.k.a., Xi Jingpin) ambitions to realise China as, again, the "Middle Kingdom". That is to say, as history should have taught us, a dictator's bent should never be discounted -- irrespective of the costs incurred to achieve it.
No country can survive without TSMC. If China did take over Taiwan, the world would have to make peace with them, because it absolutely must have those chips. And by the way, the takeover might be more like Crimea than like Poland in WWII. China just takes over without a shot fired after a campaign of increasing pressure and assuming the result. Alternatively, they might have to take out Taiwan's air defenses/air power first. Then they can flood the zone. How much damage will the Taiwanese be willing to accept to avoid their fate? (No idea.)
Thanks for the wonderful piece. I think pushing the economics a step further is important: if TSMC suffers via a loss of intermediates or a loss of customers, the company shrivels. If that happens, the staff become poachable, and throwing money at them helps move the expertise to the USA. You might need Walter White to build the fab, but if TSMC shrivels, Walter White becomes available.
All the processing equipment in the fabs comes from the US, Europe, or Japan. Each equipment manufacturer keeps a small contingent of home country engineers overseas, on rotation, to do any serious maintenance on the equipment. Without those overseas engineers, the overseas equipment knowledge, and overseas parts, none of the processing equipment continues to run. When the equiment cant be maintained, the fab cant produce product in the long term. My guess is the long term is on the order of 100 days.
Also, like there is no shortcut to go to the next node, there is no shortcut to build the next generation of processing equipment. The process machine knowledge has built up over generations of engineers and has been refined through feedback at each node of in-fab operation.
Semiconductors are just an example of the general point that it takes a world of specialized knowledge and inputs to produce any high technology product, as you suggested with your Xilinx example. Like it or not, no country is going to be able to have both high technology and independence. We are all going to have to get along, or go without.
China does not need to kinetically invade Taiwan to bring it to heel.
It can simply assert its UN-recognized rights to dominion over the island (acknowledged in Taiwan's constitution), close the Strait, and assume control of all air and sea traffic into and around the island province.
Unless the US is willing to lose several West Coast cities in a hurry, there is nothing it can do to stop China, whose military is far more powerful in the region. Nor is are Taiwanese eager to sacrifice their lives defending a corrupt backwater when they can make better money under the PRC.
After that, it's a matter of bargaining, and the Chinese are pretty good at that.
Nice article. One thing though. There is only one unique critical supplier: Dutch ASML for EUV. For all other equipment there are several suppliers. Applied Materials is by no means unique. ASMI And Be equipment are just as good. The US only has leverage by pressuring others.
A Chinese invasion of TW would destroy TSMC - the short term impact of the loss of that chip supply would be devastating for the world economy, but disproportionately felt by the Chinese who will have no alternatives other than SMIC (which will be cut off from SME from the West), while the US can still source from Samsung / Intel (and very likely will provide very heavy pressure on Samsung not to supply China). Medium-term the US is in a better position to stand up to new foundries in the US to produce those chips (Intel, possible Samsung foundries in the US), while it will take China much longer as they will have to replicate the entire semi value chain AND catch up the head start that TSMC/Samsung/Intel has over SMIC.
I also can't see how there is any realistic possibility that China either captures TSMC intact, with no damage, with no loss of IP and knowledge from critical employees and vendors. Or how China has even a shot at coercing TSMC to supply China and not the US (as you stated, all the SME inputs other critical technologies into TSMC are controlled by the US)
China's only route to semiconductor independence is the long route - the investing of hundreds of billions of dollars over a decade or more and reproduce every step of the chain while trying to catch up on the tech frontier. Which would argue against China initiating any attack on Taiwan except in the most desperate of circumstances.
Hello, sir, you may have overlooked a problem: Taiwan has been China's territory since ancient times, but Taiwan has been colonized by Japan for a period of time, resulting in China's reunification. So it is very inappropriate for you to use the word "invasion". Thank you.
Cadance Design Software is another US controlled kingpin
I just want to know which US Semiconductors stocks to invest into in case of a Chinese invasion of Taïwan. When TMSC goes offline, US chipmakers will pickup the slack and their stock price will go up. Please tell us that. Where to invest our money ?
What about the option of the US destroying say 4 leading edge TSMC fabs in Taiwan, as soon as they "discover" the CCP are invading? Disturbing the leading edge development of Semis for say 2 years, but ensuring the CCP does not hold the reigns to TSMC, and others picking up the slack.
CCP could then focus on all the human and natural treasures that Taiwan has to offer them...
You don't think this is a realistic scenario?
So what if for 2 years we can't make better graphic chips? Have to make cars where the seatbelt doesn't beep? Life goes on. Growth slows- ALOT- ok. CCP however NOT running the tables on the semis front...
I would think it would be possible to get most of the TSMC workers out of Taiwan by plane or boat if China invades. With all the equipment being produced in the west they could set up shop in the west. Fabs can be built in advance so they can move into ready facilities.
Taiwan and China are already tightly coupled. Taiwan is represented in China's Congress, for example, and its freighters uniquely enjoys the right to sail on China's vast internal waterways–giving it huge time and cost advantages over rivals. President Xi has relatives in Taiwan and the Taiwanese trust him more than they trust their own politicians.
My Chinese friends laugh at the idea of a purely military takeover of Taiwan. More likely, they say, is a bland official announcement, "As of midnight tonight, all movement of goods and persons to and from Taiwan must clear PRC Customs and Immigration. Otherwise, no changes are contemplated".
Of course, this would be enforced by China's immensely powerful military and a picket line of missile patrol boats around the new ADIZ. The Taiwan Strait would, a fortiori, become domestic waters, closed to all military traffic.
In other words, business as usual, except that global power would shift decisively in China's favor both through its control of IC imports and exports, and the impotence of the US military to hinder the move.
Question: I get your point, and agree, that tools suppliers like ASM Litho, KLA- Tencor, Applied Materials and Lam Research are critical to any fabs ability to operate at the cutting edge. Just curious, where do the US based design software companies Synopsys and Cadence fit in on all this? Judging from their incredible margins and my understanding of the business it is basically impossible to design leading edge, or maybe any, chips without their software.
This article seems to be based on the dubious assumption that China wants to invade Taiwan. China wants to gain control over Taiwan in the long run and invading it is not the most convenient way to achieve such control. Constant military pressure with minor clashes is Chinese optimal strategy. If there will be war, then China will destroy large part of Taiwanese military in one day without invading Taiwan. The fact that Taiwan can not defend against China will become completely obvious in the coming years and China will slowly establish political control over Taiwan.
OK in parts, but some really unforgiveable gaffes. The author at one point makes the correct point TSMC can't function without continuous inputs from US companies and those that must follow US dictates, like Japanese suppliers of semicon raw materials. The US can shut down all production, yes all production, at TSMC at any point and the fabs in Taiwan would no longer be able to operate. Not a problem money can solve.