China's chip-making ambitions face setbacks Trade restrictions, sanctions, and other challenges are putting a dampener on China's ambition to become a chip manufacturing hot spot. Research firm IC Insights is projecting the Chinese semiconductor foundry market share to remain flat through 2026, while rival manufacturers in the US, Taiwan, Korea, and other countries grow …
ASML's EUV scanners are some of the most complex and sensitive machines made by man, with over 100,000 parts many of which must be maintained to sub micron tolerances. It takes SIX MONTHS to install and calibrate one after delivery, so "safely transporting" one without months of advance warning is simply not possible.
Without full time support from ASML, they are basically useless. There is constant tuning and calibration that has to happen, this isn't something you can install and have operate for months with yearly service like a car. TSMC engineers could walk away in the middle of a shift, no need to blow anything up.
Chinese engineers could take them apart to see how they're made, but so many parts are one offs that are either made by ASML or from a vendor that is likewise not selling to China (and that list would get a lot longer if they invaded Taiwan) so it would take them years to clone it with no guarantee of success. And meanwhile the state of the art would have progressed well beyond that point so China would be as far behind as they are today.
If China wants to catch up, they'll need to do it the hard way, by inventing their own - including the hundreds of subcontractors that supply various parts and materials used to make them.
ASML's machines aren't a box, they're more like a chunk of process that comes with their own set of engineers. My daughter had a next door neighbor that worked at an Intel plant but didn't work for Intel, he worked for ASML managing a team that kept the kit running and in calibration.
The light source they use in this lithography illustrates the kind of problems they have with this technology. It comes from bombarding droplets of molten tin with an electron beam, that in itself is taking the term 'extreme' a bit literally. I can't even begin to imagine how mask registration is managed.
The real danger to 'us' is that 'they' aren't developing technology from scratch. They might copy what they can but they're not going to have the mistakes and compromises that came with the first products. So when -- and it really will be 'when' -- the Chinese develop their version of this technology its likely that it will be highly competitive. (...and put any thoughts of them being Chinese etc. out of your mind -- politics aside the Tiawanese are also Chinese.)
China might be able to develop their own capabilities to an extent now, but you would have to be seriously deluded as a western firm to open up any sort of manufacturing plant (semiconductor or frankly anything) in China these days. Hell producing anything apart from the dumbest of dumb items in China these days is a guarantee that you will find yourself undercut by a Chinese firm who have basically just stolen your design (and probably manufactured it on your own production line as well).
That lack of external investment will eventually see China suffering, but it is rather insane that it's taken this long for Western companies to cotton on to this fact...
Why would a western company open a manufacturing facility in China? To get access to the Chinese market. If you don't make stuff there, China may cut off your imports. Or your own government may cut off your exports. Of course China may throw you out and force you to sell your factory -- probably at a loss. But likely they won't.
China's economy is currently about the same size as the US economy and will in all likelihood grow faster. It certainly has grown faster for the past four decades. There are a LOT of Chinese. Four times as many as in the US. Twice as many as in Europe. And many of them lack things those of us in developed countries take for granted. They're likely going to be buying a lot of stuff in the next few decades.
So yes, making stuff in China is risky. But there's money to be made ... probably. And isn't making money the object of business?
Sure, but some things simply can't be made there when western companies aren't allowed to bring ASML EUV scanners into China (and if one surreptitiously did so, without installation, calibration, and round the clock maintenance from ASML personnel, they are utterly useless)
Might take a while. China can make acceptable versions of most things including complex things like nuclear reactors. But there are a few areas where they're tried and haven't quite succeeded -- high performance jet engines for example. And all the big players in the semiconductor business seem to be running as fast as they can to make the latest cutting edge products. It's not like the current leaders are sitting around drinking rum drinks with little umbrellas while congratulating themselves on another great sales year. China may well need to run as fast as they can just to stay a few years behind.
On the other hand, if you figure that China is mostly interested in making and selling consumer goods, do most of those really require cutting edge? One suspects that the chips, if any, in your washing machine or electric toothbrush are quite a ways behind state of the art (and cost about $0.07 each in quantity 10,000). China can presumably turn out as many of those as the markets need.
A point you forget to address is, what if other simpler methods of manufacturing at nano scale are discovered. For instance, nano-imprint is a method the Japanese are pursuing. It is altogether a simpler process than EUV manufacturing.
It is important to bear in mind that EUV manufacturing is not the end goal, nano scale engineering is the objective, irrespective of whether you achieve it with EUV or some other method.
Researchers have put forward numerous methods, such as adopting a different wafer material like carbon nano tubes. Carbon nano tubes, researchers believe are a better ideal medium for ICs or transistors.
A breakthrough in these areas can render all the investments in EUV redundant, and talking about investments, there is the presumption that everything technology sells. This was certainly not the case in the 80s and early 90s, before the internet age, computers to most people were glorified type writers.
Today most commercial advanced node manufacturing is directed towards producing ICs for mobile phones. It is certainly my personal contention that the market is saturated. Like most people I find no need to change my personal mobile phone annually to keep TSMCs fabs humming and Apple richer.
And as we approach more inflationary times (covid and Russian sanctions) with a slow down in credit, and growth of bad debts people are becoming uncertain about the future. Mobile phone expenditure is discretionary i.e. not important or necessary to households, demand for the devices could collapse.
If that happens, how then are these semiconductor fabs going to pay for their investments in EUV? The subsequent result would be the mothballing of capacity leading to a secondary market in semiconductor equipment, the latter would be disastrous for ASML.
These scenarios are not absurd or unlikely they are very real possibilities.
ASML was only able to make EUV scanners because of all the companies they relied upon to make parts and materials to make those parts to their spec. ASML could have started on day one with the blueprint to the finished product and it would have still taken nearly as long to reach that end goal because so many other companies needed to invent new technologies or refine them to finer tolerances or purity than had previously been achieved.
It isn't that China is incapable of doing it, it is that the combined engineering might of pretty much the whole world worked on those subproblems for the better part of two decades until ASML succeeded in putting everything together into something that not only worked in a lab, but in a factory at high enough volume to handle customers who require hundreds of millions of chips a year like Apple.
China's economy is currently about the same size as the US economy
No it isn't. 2020 US GDP was a shade under $21 trillion. China's $14.7tr. That's with a population 4 times the size of the US. While China has faster growth rates, they are steadily falling. And a smaller percentage of a large pie is still a very large number.
China might still catch up and overtake. But that requires the Communist Party of China not to fuck things up. And I don't think that's guaranteed.
2021 figures are in China has an $18 trillion GDP economy https://www.scmp.com/economy/china-economy/article/3163720/china-gdp-growth-2021-beats-expectations-narrowing-gap-us
While the US has remained flat due to covid. In short they're a couple of years from overtaking.
I lazily took Google's first result last time. Just checked on 2021 figures and it's now China 17.7 USA 22.99. So still a way to go. And China's growth rates are dropping and it's just beginning to suffer from the same sort of demographic problems of the developed economies in Europe, but without being a developed economy yet.
There's still a good chance that China's economy will overtake the US soon - but I don't think it's a certainty. The US is still very productive and manages higher growth than the developed ecomomies in Europe do.
The question is also whether we're going to go back to globalisation-as-usual. Or whether the supply shocks of Covid coupled with the political shocks of China and particularly Russia's more aggressive foreign policy lead the developed economies to get a bit more self-protective. The US have been trying to for the last couple of years, if not in any organised way. But Europe has been talking about strategic equidistance between them and the US and how more trade will lead China to westernise more and become less aggressive. But Germany just entirely re-wrote their last 50 years of foreign policy last week, over Russia's invasion of Ukraine - so I don't think it's a certainty that foreign investment in China is going to continue at the previous pace. Even if our economies don't try to bring work back home, we might try to diversify our outsourcing to places like Vietnam, India and Indonesia. To cut dependence on China. Or maybe European politicians will just go back to sleep. I think it would be a fool that made firm predictions.
(...and a Deep Purple Album)
But the Japanese progressed. I wouldn't write off China quite so easily.
NYT: 'Made in Japan' (Without the Inferiority Complex)
Japan`s Quality improvement Process, Total Quality Control and employee engagement
The rise of Japan: How the car industry was won
Japan lacks natural resources, and so had to find other means to add value to products. China can't build cutting edge fabs overnight: but can incrementally improve what they have already. They are in catch-up mode, but Japan shows catching up and surpassing is possible.
The difference is that China is not currently run by a friendly government. It's also a very autocratic government. This could have several effects. It could reduce China's ability to develop economically by fucking things up. There are some signs of that already. There's also the problem of dictatorships tolerating their friends stealing loads of the money, which if allowed to get to Russian levels destroys future investment in the economy.
Next you have the problem that if you keep pissing off all the foreigners, they might gang up on you. And given what's just happened with Russia, and the fact that the US had already started to build mechanisms to stop technology transfer - this again could limit China's potential growth. And it's not just the US, but Japan, Korea and Taiwan, Australia, the UK and the EU may also join the party after their experiences with Russia.
Also if Xi Xinping continues on his current path, he's making China less of an attractive place to live. Already lots of bright Chinese people go abroad to study, but many of them don't bring that knowledge back to China, since they prefer to live somewhere less repressive. And he may also drive out potential foreign investment.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
