Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/corruption-ineptitude-chinas-chip-failure.16522/page-2
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Corruption, Ineptitude, China's Chip Failure

Solar panels - China, far and away
Nuclear Power Plants - Russia, France is up there
EV Batteries - China
EV vehicle production - China, by a lot
AI SW & HW - Hmmm... China has had some impressive results, but so has the US and Israel
Memory production - Korea
Ship building - definitely Korea
Smart phones - China for mechanical assembly, the US and Korea for design
5G - Unclear, though IP in the standards group points to China. Which is a bit amusing, considering China's policies regarding other country's IP
Space exploration - US, for decades, though US manned programs are becoming less important (Which I agree with. Human space travel is silly.)
Car manufacturing - by what measure? In production volume, China
Supercomputers - China is arguably a leader, though only for internal use

So 5/12, 6/12 if we count the auto industry by volume. Did I get anything woefully incorrect?

Some times massive subsidies work, some times don't.

In the mainland, you could've gotten 3 million USD just for having word "solar" in the name of the company around a decade ago. Of course, the level of graft, and abuse was immense. But in the end, most survivors after the party was over managed to dominate the market.

BOE took decades living off government subsidies to eventually become self-sufficient.

On other hand — flash, DRAM, mid-tier semi fabricators who lived off subsidies ALL failed.
 
China's failure to develop a state-of-the-art chip industry failed due to corruption and ineptitude combined with a lack of controls and planning. Almost any time, there are large amounts of money, corruption, and inefficiency rules, just like US medical provides lower quality medical care at the world's highest prices. Any government project is hard to run for many see a bottomless pit of money, and having worked on many billion-dollar projects, the US is no exception, but better than most due to a truly free press combined with the greatest Constitution and Bill of Rights. In any dictatorship, fear rules, discouraging free thinking, which is the key to true progress. SemiWiki is but one example of this.

Is it legal or at least fair to subsidize the commercial companies by the government? I mean a government in free commercial world can support R&D activities in companies but is it OK for providing money for production investment to the companies that are targeting profit from their commercial activities? Isn't this against the commercial agreements/MOUs between the governments in the world? Can a government distribute taxpayers' money to commercial companies in order to make them better compete agains other companies that are using the funds created by themselves from their commercial activities? Is this fair? What is the difference then between PRC and the USA?
 
Is it legal or at least fair to subsidize the commercial companies by the government? I mean a government in free commercial world can support R&D activities in companies but is it OK for providing money for production investment to the companies that are targeting profit from their commercial activities? Isn't this against the commercial agreements/MOUs between the governments in the world? Can a government distribute taxpayers' money to commercial companies in order to make them better compete agains other companies that are using the funds created by themselves from their commercial activities? Is this fair? What is the difference then between PRC and the USA?
Of course this is possible - and indeed happens to a varying extent everywhere (Boeing arguably did this when competing against Airbus who did much the same, but though different 'workarounds'). State subsidies are officially illegal in the EU - but the French still seem to find ways round this.

The real question is "is it wise and sustainable" ? The problem with subsidies is always the opportunity cost - what else might have been achieved with the same money and resources and whether that is better for the country as a whole than "protecting jobs" at a cost of [let's say - and this is not unknown] $1m/employee.

If a competing country is spectacularly mis-allocating resources and thereby becoming less competitive, do you actually need to complain ("never interrupt an opponent when he is making a mistake" as Napoleon said).
 
I think China failed less than the Western EQP supplies succeeded. Here’s how.

A fab tool has a remarkable accumulation of mechanical and software know-how embedded, and the cost reflects that investment. China’s business model is to reverse engineer and cut the mechanical and software cost with local suppliers.

This has been tried before. China’s model was Korea’s and Taiwan’s model. And so far, there are few Korean and Taiwanese EQP vendors. This should tell you something—for decades, AMAT and others have been building enough mechanical and software traps to prevent effective reverse engineering.

The way these traps work is pretty simple. Take an off-the-shelf part, customize it, then set up exclusive supply with the vendor. Keep track of the tool numbers and only sell that part to legitimate, verified tool owners. China (or Taiwan, or Korea) have no way to penetrate this, that I know of. And they try, constantly.
 
I think China failed less than the Western EQP supplies succeeded. Here’s how.

A fab tool has a remarkable accumulation of mechanical and software know-how embedded, and the cost reflects that investment. China’s business model is to reverse engineer and cut the mechanical and software cost with local suppliers.

This has been tried before. China’s model was Korea’s and Taiwan’s model. And so far, there are few Korean and Taiwanese EQP vendors. This should tell you something—for decades, AMAT and others have been building enough mechanical and software traps to prevent effective reverse engineering.

The way these traps work is pretty simple. Take an off-the-shelf part, customize it, then set up exclusive supply with the vendor. Keep track of the tool numbers and only sell that part to legitimate, verified tool owners. China (or Taiwan, or Korea) have no way to penetrate this, that I know of. And they try, constantly.

Actually,the speed of adopting domestic EQP in Chinese fabs is unprecedented

From publicly gathered bidding infos,in 2022 Q1 and Q2,the accumulative domesticalize rate of fab tools is 35%. Two years ago this rate was around 5%.


2022-08-13_142819.png

2022-08-13_142753.png
 
Is it legal or at least fair to subsidize the commercial companies by the government? I mean a government in free commercial world can support R&D activities in companies but is it OK for providing money for production investment to the companies that are targeting profit from their commercial activities? Isn't this against the commercial agreements/MOUs between the governments in the world? Can a government distribute taxpayers' money to commercial companies in order to make them better compete agains other companies that are using the funds created by themselves from their commercial activities? Is this fair? What is the difference then between PRC and the USA?
Is it fair, not really. But we don't live in a world with free markets. The EU heavily subsidies what little industry they have left and use massive trade barriers. PRC does the same thing, as does Brazil, Japan, Korea, the ROC, and the US within certain sectors. Regardless of the opinions of us or economist, the rise of the semiconductor industry within the ROC and ROK was done with heavy government investments in R&D/education. Additionally there was a large amount of corporate welfare to build these companies up to the capacities/scale they needed to become profitable (heck if memory serves SK Hynix needed to get propped up at some point because they couldn't pay off their debts). So unfortunately it seems like if the US and EU want to increase their manufacturing/assembly share then they will have to do it via artificial means like the ROK and ROC did many years ago.
 
Actually,the speed of adopting domestic EQP in Chinese fabs is unprecedented

From publicly gathered bidding infos,in 2022 Q1 and Q2,the accumulative domesticalize rate of fab tools is 35%. Two years ago this rate was around 5%.


View attachment 861
View attachment 862
This is good progress, and should be applauded. There is a but coming…

Taiwan, Korea, and the US also have local suppliers for about 35% of the fab equipment. This is the first tier of supplier, not very value-added, basically you bid for it and they award it. The performance of this equipment is not critical to chip quality.

In the next layer, there is no competition, it’s 100% consolidated around a few global vendors in USA, Japan and Netherlands, I hope China cracks this nut. Honestly. It would be better for everyone, AMAT, Lam, ASML, TEL, KLA included, to have competition again.

I see the headwinds, including political embargos, so I’m rooting for China (and competition, and free market capitalism) but think it ain’t happening.
 
Last edited:
This is good progress, and should be applauded. There is a but coming…

Taiwan, Korea, and the US also have local suppliers for about 35% of the fab equipment.

The difference is incentives. Fabs in KR/TW/US have no risk of being cutoff EQP supplies,so they are not keen on adopting domestic EQP. Where as Chinese fabs facing ever increasing difficultly of acquiring EQPs from overseas,as the US adding more restrictions everyday. So they have no choice but adopt domestic EQPs as much as possible and ASAP.
 
Back
Top