Regarding building wafer fabs in Japan vs. the US, I have no reason to doubt your assertions about the regulatory climate in Japan being more favorable. You've convinced me that you know more about Japan than I. However, I must reserve judgement whether their bureaucracy is better than ours.
No matter. My comment has to do with your thesis that Japan is a solution for Taiwan's existential crisis. I'm reminded of the three little pigs. The straw house (Taiwan) offers no protection, but the wooden house (Japan) does. Your opinion is that building a brick house (US) is neither feasible nor cost effective.
The real issue is whether moving a strategic trans-Pacific supply line from Taiwan to Japan makes any difference, when the real peril is that the Pacific shows signs of becoming a Chinese lake. A slight overstatement, but one that China would endorse. As another reader has pointed out, Japan is worringly analogous to Finland.
After having spent my career in the semiconductor industry in the US and being familiar with the industries in Europe, Taiwan, Korea and Japan, I don't see a compelling argument that the US is uncompetitive in engineering skills, manufacturing equipment, semiconductor technology or supply infrastructure. Ditto for the large labor pool of fab operators and maintenance technicians. Japan, by comparison, is faced with shrinkage of the key labor demographics that drive this industry. Finally, the US is the World's engineering and physics classroom.
Japan is very different to Finland in terms of its size vs the giant next door and in terms of access. Finland is (was) 5% or less of the Soviet population. Japan is a good 10% of the PRC population.
Furthermore Finland (as it is noticing again recently) has a very long land border with Russia. Japan has a sea. The only realistic way to invade Japan is via S Korea (which in turn is likely to object strenuously).
My experience (I live in Japan) of Japanese bureaucracy is that it tends to take a while to get going and then it can be extremely efficient as long as you/it stay on the rails. Also in many ways Japan is less bureaucratic than the US (well less than California or NY, I can't talk about places like Wyoming). Zoning is way simpler. Permits are generally granted assuming you pass all the required seismic etc. hurdles regarding building safety. Once you get the permits building tends to be very quick. For example, Murata (a electronics component manufacturer) has finished two extensions to existing plants in Japan this year where construction started in 2022.
Another subtle downside is interference from unproductive labor unions (supported by government) in both the US and Canada. A new fab plant in Arizona is getting some union blowback about hiring.
Yes, there's bad management and bad unions. But the incentives in US labor law make the latter 10x worse.
In the '80s, I grew up with extended family members in the US domestic auto sector, both management and labor. The myopia and entitlement mindset within both was stunning to a teenager told to always "work hard". As was the subtle racism directed at Japanese car brands.
It's better today, but there's still a large productivity gap between union and non union plants.
I think Canon’s cheaper chip fabrication will feed
a growing need. Given the real-world implications of AI, I think high-end chip production will need to double. The transformation (pun intended) of chips is an inflection point. Eventually, there will be niches for bespoke chips at scale that will fill. The one-size-fits-all chip won’t last, which is good for developers. Also, if the industry wants to claim it’s “green,” it will need to put fabs near large fresh-water sources and hydro- or solar-energy sources.
Japan has a lot of fresh water. Sometimes it has far too much and floods. Rainfall (snowfall) is regular and there are already a lot of dams, some for hydro, others for flood prevention. Both can absolutely be used as sources of water for industry
Doesn’t this mean they have to get strong seismic stabilization even in places that don’t have many major earthquakes? How much more expensive is it to stabilize in these places, if they have to do it everywhere?
I assume they engineer this into their fabrication plants, especially since they chose Kumamoto which had a big earthquake in 2016. I visited Kumamoto castle last year and it had sustained significant damage from that quake, but is mostly still open for tours. When the castle was rebuilt (in 1960 and again in 1998-2008, as they often are there) it had earthquake reinforced cables and dampers installed inside which are not visible externally, but a section of the older unreinforced moat wall collapsed. Bullet trains also can be dangerous in earthquakes, but the Shinkansen has seismic sensors which stop trains in seconds, and there hasn’t been an earthquake caused accident yet.
Good article and a lot of valid points. Three factors on the other side. First, the language of business is English... except in Japan. China/South Korea...pretty much anywhere else in the world... it is possible to converse on business/technology in English. The second factor is culture... Japanese culture has great qualities, but it is quite a contrast to western culture....especially around role of individuality. Finally, the historical relationships between Japan and its neighbors (China, South Korea, Philippines..) is still a bit raw at the cultural level. US companies routinely fly support staff from mainland USA vs their Japanese operations because of the historical issues.
All of these factors create very real headwinds for deep partnerships. However, it does happen..and the newer generations are much more connected to the external world and the memory of Imperial Japan is fading... so who knows.
I’ve worked for joint Japanese/American tech companies and these issues were not factors. High level managers are usually competent in English conversation and written English, and Japanese weakness in English compared to Korea and Taiwan is more in conversation rather than reading or writing. Lack of emphasis on individualism is not a problem for large corporations as opposed to startups, but actually a positive factor. The only real cultural violence in Asia against Japanese people and companies that I’ve seen is in China and mostly when prodded by the PRC (see the ridiculous bans on seafood imports because of the release of completely safe sea water in Fukushima). In September I was in Kyushu and there were many tourists mainly from South Korea and Taiwan and no cultural hostility, and Japanese also regularly travel to those countries.
Buzen, You may well be right as your experience in Japan appears to be more recent. This was not my experience from about a decade ago working broadly across the world (including Japan). As a senior executive for a US company, literally the only place in the world where English was not the language of discourse was Japan. I don't mean chatting in conversational items...I mean a real discussion on complex topics. Now. things may have changed......this would be goodness for Japan and the world in general.
More recently, I starting working with the US sub for a large Japanese company, and I saw similar issues. Perhaps this is more an indication of large Japanese conglomerates vs the broader business landscape.
I was about to ask why not South Korea because nominal GDP per capita is lower there (probably meaning lower labor costs) and there's already a semiconductor scene there. Furthermore, the country's young people are the most likely to hold a college degree in the world with 68% of 25-34 year olds holding a bachelors degree.
Though on the flipside, South Korea has been catching up to Japan in nominal GDP per capita (making labor costs higher in the long run) and as of 2023, is only a hair under Japan, making it all caught up. And with Japan's currency becoming weaker, I guess labor costs are now officially lower in Japan than in SK. Also, 59% of 25-34 year olds in Japan have a bachelors degree.
So maybe Japan is a better long term choice than SK.
“Now Japan’s government finally realizes that the country needs to be a collection of valuable nodes in a production system that stretches all across Asia, Europe, and North America. And that means attracting FDI.”
FDI into China has dwindled. Money has to go somewhere, most likely into the moat of Southeast Asian countries surrounding China. Geographic diversification of chip fabs is an economic weapon China doesn’t have in the long-term big-picture. Xi’s blustering and bullying has caused long-term high-tech damage to China.
As a long-term investor in ASML (2013), it would be a mistake to underestimate Cannon and its CEO. If I were looking for another picks-and-shovels investment in the chip sector, Cannon would be my choice. Japan can compete with anybody and needs an electronics Renaissance.
I had very productive collaborations within IBM Japan in the 1980s and early 1990s, then turned in other directions. What happened? Nowhere in your article do you mention Fujitsu or Hitachi, except as a source for poaching the remaining signs of life.
This is overall good. But we need to also be doing everything possible to bring more manufacturing to the continental USA. Not just for economic reasons, but because too much is concentrated in an area prone to big earthquakes and other geologically localized potential problems between Taiwan and Japan.
Why did Japan get left behind by ROK and Taiwan? One argument I’ve heard is that, despite Japan’s massive financing potential, its companies struggle with the enormous capital investments (and risk) that goes into doing huge fab projects.
But why? I think there are two related and interesting issues here. One is Japan’s risk appetite, which won’t be easy to change, especially with an aging population. It might go in the wrong direction.
Second is the role of trading houses. I have no doubt that Japan’s great companies would be much greater if they didn’t compete for talent, capital and policy influence with trading houses.
Noah, would love to see your detailed take on the role of trading houses.
Regarding building wafer fabs in Japan vs. the US, I have no reason to doubt your assertions about the regulatory climate in Japan being more favorable. You've convinced me that you know more about Japan than I. However, I must reserve judgement whether their bureaucracy is better than ours.
No matter. My comment has to do with your thesis that Japan is a solution for Taiwan's existential crisis. I'm reminded of the three little pigs. The straw house (Taiwan) offers no protection, but the wooden house (Japan) does. Your opinion is that building a brick house (US) is neither feasible nor cost effective.
The real issue is whether moving a strategic trans-Pacific supply line from Taiwan to Japan makes any difference, when the real peril is that the Pacific shows signs of becoming a Chinese lake. A slight overstatement, but one that China would endorse. As another reader has pointed out, Japan is worringly analogous to Finland.
After having spent my career in the semiconductor industry in the US and being familiar with the industries in Europe, Taiwan, Korea and Japan, I don't see a compelling argument that the US is uncompetitive in engineering skills, manufacturing equipment, semiconductor technology or supply infrastructure. Ditto for the large labor pool of fab operators and maintenance technicians. Japan, by comparison, is faced with shrinkage of the key labor demographics that drive this industry. Finally, the US is the World's engineering and physics classroom.
Japan is very different to Finland in terms of its size vs the giant next door and in terms of access. Finland is (was) 5% or less of the Soviet population. Japan is a good 10% of the PRC population.
Furthermore Finland (as it is noticing again recently) has a very long land border with Russia. Japan has a sea. The only realistic way to invade Japan is via S Korea (which in turn is likely to object strenuously).
My experience (I live in Japan) of Japanese bureaucracy is that it tends to take a while to get going and then it can be extremely efficient as long as you/it stay on the rails. Also in many ways Japan is less bureaucratic than the US (well less than California or NY, I can't talk about places like Wyoming). Zoning is way simpler. Permits are generally granted assuming you pass all the required seismic etc. hurdles regarding building safety. Once you get the permits building tends to be very quick. For example, Murata (a electronics component manufacturer) has finished two extensions to existing plants in Japan this year where construction started in 2022.
https://corporate.murata.com/en-us/newsroom/news/company/general/2023/0427
https://corporate.murata.com/en-us/newsroom/news/company/general/2023/0913
Another subtle downside is interference from unproductive labor unions (supported by government) in both the US and Canada. A new fab plant in Arizona is getting some union blowback about hiring.
Yes, there's bad management and bad unions. But the incentives in US labor law make the latter 10x worse.
In the '80s, I grew up with extended family members in the US domestic auto sector, both management and labor. The myopia and entitlement mindset within both was stunning to a teenager told to always "work hard". As was the subtle racism directed at Japanese car brands.
It's better today, but there's still a large productivity gap between union and non union plants.
I don’t know if Noah is right, but I hope so. Japan is a pretty wonderful place and it would be nice to see it regain its mojo. Tennoo heika banzai!
There are 3 great non-military Optics companies:
Zeiss - they invented optics
Canon
Nikon
Nikons optical design and engineering skills are all the equal to Zeiss. Especially in optical lithography stacks.
Additionally, the most sophisticated mechanical and thermo-mechanical engineering is an absolute requirement. Zeiss and Nikon have these.
The 1st cousin to optical lithography with somewhat related genealogy is reconnaissance optics, but lithography has surpassed that technology.
I think Canon’s cheaper chip fabrication will feed
a growing need. Given the real-world implications of AI, I think high-end chip production will need to double. The transformation (pun intended) of chips is an inflection point. Eventually, there will be niches for bespoke chips at scale that will fill. The one-size-fits-all chip won’t last, which is good for developers. Also, if the industry wants to claim it’s “green,” it will need to put fabs near large fresh-water sources and hydro- or solar-energy sources.
Japan has a lot of fresh water. Sometimes it has far too much and floods. Rainfall (snowfall) is regular and there are already a lot of dams, some for hydro, others for flood prevention. Both can absolutely be used as sources of water for industry
One reason why they shouldn't: earthquakes. Even small shakes ruin the smallest fabrications. Taiwan isn't great for this, but Japan is even worse.
Doesn’t this mean they have to get strong seismic stabilization even in places that don’t have many major earthquakes? How much more expensive is it to stabilize in these places, if they have to do it everywhere?
no one is better at building infrastructure w seismic activity in mind than them
I assume they engineer this into their fabrication plants, especially since they chose Kumamoto which had a big earthquake in 2016. I visited Kumamoto castle last year and it had sustained significant damage from that quake, but is mostly still open for tours. When the castle was rebuilt (in 1960 and again in 1998-2008, as they often are there) it had earthquake reinforced cables and dampers installed inside which are not visible externally, but a section of the older unreinforced moat wall collapsed. Bullet trains also can be dangerous in earthquakes, but the Shinkansen has seismic sensors which stop trains in seconds, and there hasn’t been an earthquake caused accident yet.
https://theculturetrip.com/asia/japan/articles/8-ways-japan-prepares-for-earthquakes
Good article and a lot of valid points. Three factors on the other side. First, the language of business is English... except in Japan. China/South Korea...pretty much anywhere else in the world... it is possible to converse on business/technology in English. The second factor is culture... Japanese culture has great qualities, but it is quite a contrast to western culture....especially around role of individuality. Finally, the historical relationships between Japan and its neighbors (China, South Korea, Philippines..) is still a bit raw at the cultural level. US companies routinely fly support staff from mainland USA vs their Japanese operations because of the historical issues.
All of these factors create very real headwinds for deep partnerships. However, it does happen..and the newer generations are much more connected to the external world and the memory of Imperial Japan is fading... so who knows.
I’ve worked for joint Japanese/American tech companies and these issues were not factors. High level managers are usually competent in English conversation and written English, and Japanese weakness in English compared to Korea and Taiwan is more in conversation rather than reading or writing. Lack of emphasis on individualism is not a problem for large corporations as opposed to startups, but actually a positive factor. The only real cultural violence in Asia against Japanese people and companies that I’ve seen is in China and mostly when prodded by the PRC (see the ridiculous bans on seafood imports because of the release of completely safe sea water in Fukushima). In September I was in Kyushu and there were many tourists mainly from South Korea and Taiwan and no cultural hostility, and Japanese also regularly travel to those countries.
Buzen, You may well be right as your experience in Japan appears to be more recent. This was not my experience from about a decade ago working broadly across the world (including Japan). As a senior executive for a US company, literally the only place in the world where English was not the language of discourse was Japan. I don't mean chatting in conversational items...I mean a real discussion on complex topics. Now. things may have changed......this would be goodness for Japan and the world in general.
More recently, I starting working with the US sub for a large Japanese company, and I saw similar issues. Perhaps this is more an indication of large Japanese conglomerates vs the broader business landscape.
BTW... here is an article which describes the current situation: https://www3.nhk.or.jp/nhkworld/en/news/backstories/2186/ ... it appears there are efforts to shift towards English, but it is not easy.
I was about to ask why not South Korea because nominal GDP per capita is lower there (probably meaning lower labor costs) and there's already a semiconductor scene there. Furthermore, the country's young people are the most likely to hold a college degree in the world with 68% of 25-34 year olds holding a bachelors degree.
Though on the flipside, South Korea has been catching up to Japan in nominal GDP per capita (making labor costs higher in the long run) and as of 2023, is only a hair under Japan, making it all caught up. And with Japan's currency becoming weaker, I guess labor costs are now officially lower in Japan than in SK. Also, 59% of 25-34 year olds in Japan have a bachelors degree.
So maybe Japan is a better long term choice than SK.
So warnings about the Japanese eating us alive were true, just 40 years early 🤔
“Now Japan’s government finally realizes that the country needs to be a collection of valuable nodes in a production system that stretches all across Asia, Europe, and North America. And that means attracting FDI.”
FDI into China has dwindled. Money has to go somewhere, most likely into the moat of Southeast Asian countries surrounding China. Geographic diversification of chip fabs is an economic weapon China doesn’t have in the long-term big-picture. Xi’s blustering and bullying has caused long-term high-tech damage to China.
As a long-term investor in ASML (2013), it would be a mistake to underestimate Cannon and its CEO. If I were looking for another picks-and-shovels investment in the chip sector, Cannon would be my choice. Japan can compete with anybody and needs an electronics Renaissance.
I had very productive collaborations within IBM Japan in the 1980s and early 1990s, then turned in other directions. What happened? Nowhere in your article do you mention Fujitsu or Hitachi, except as a source for poaching the remaining signs of life.
When I first started working in the Semiconductor Equipment Industry, in the early 90s, Japan was the dominant player in the global market with 50% share of Semiconductor Revenue. See the Siemens blog here https://blogs.sw.siemens.com/cicv/2023/12/07/the-resurgence-of-japans-semiconductor-industry/
This is overall good. But we need to also be doing everything possible to bring more manufacturing to the continental USA. Not just for economic reasons, but because too much is concentrated in an area prone to big earthquakes and other geologically localized potential problems between Taiwan and Japan.
Why did Japan get left behind by ROK and Taiwan? One argument I’ve heard is that, despite Japan’s massive financing potential, its companies struggle with the enormous capital investments (and risk) that goes into doing huge fab projects.
But why? I think there are two related and interesting issues here. One is Japan’s risk appetite, which won’t be easy to change, especially with an aging population. It might go in the wrong direction.
Second is the role of trading houses. I have no doubt that Japan’s great companies would be much greater if they didn’t compete for talent, capital and policy influence with trading houses.
Noah, would love to see your detailed take on the role of trading houses.
Related question: what do you recommend reading for the Japanese government/Emperor perspective on entering WWII and the empire growth that led to it.
We have NEPA to blame but it helps species stay off ESA lists. That saves fortunes.