I did not see any proof for any of those statements nor is he considered a known expert in the field. Meaning that I could write a 180 of that article and it would hold just as much weights as that "commentary".
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
I think you are software engineer so you could easily say that you're simply a "down stream end user evaluator". On the other hand he's a political science major...
Chinese semiconductors are attracting pundits like months into flame. I remember there was an Australian sociologist that wrote about Chinese semiconductors in Foreign Policy lmao. Anyway, these days he's an Indian simp.
Lol imagine if it were the other way around and China dominated semiconductor production and was weak in software and their ecosystems. I wonder what the Foreign Policy articles would be like then.
PhD in Politics at Princeton University, PD Soros Fellow.
Great. Another overpay National Security "Expert".
Maybe, but they have SMIC, the 5th largest in the world.
The Chinese goverment could had enforced the localization of their tools-materials supply chain sooner, like enforcing the use of SMEE frontend scanners sooner at the expense of ASML to create a virtuous cycle, at least in the mature nodes-mems and when goverment money and subsidies was involved but I think the biggest complainer of that would have been the US and some Western countries. "China bad for not giving a fair chance to US toolmakers over domestic ones"
There is always risks involved with this kind investment, be sure that US will find out that, but unwillingness to take that risk would lead any company and country to nowhere. The free market may work best after the gears already started spinning.
Maybe the biggest mistake of Chinese companies was putting their trust in unreliable US companies and Chinese goverment not enforcing the localization of the supply chain, at least of mature nodes sooner.
Starting with "low" yields, like 70%, according to an "expert" in Weibo in a domesticated line is an start and could be solved sooner with hard work. But they have to start building localized fabs to start solving issues. The Chinese goverment could subsidize the "low" yield until improvements are made.
For now the localization of the semiconductor industry getting higher than ever before, with domestic toolmakers having record sales, lot of tools-materials projects being develop and in the process of verification. More research going to tools, materials and software. Manufacturing processes being redesigned for domestic tools. New process being designed around domestic tools. It could be that the glass is half full, in fact is a bit more than half full.
I don't think they need have to reach 100% independent because as the domestic supply industry get stronger the pushback against export controls from companies and US allies will also grow stronger. Because one of the reasons why US is doing this is because a group overpay think tanks pundits like this ... guy keep pointing out the "perceived" low market shared of China own domestic companies as a weakness to be exploited.
For sure there have been frauds in China with some of those projects like Wuhan Hongxin and and flops like Tsinghua's various memory projects outside of YMTC. We shouldn't deny those, and they should be highlighted as lessons learned. When you throw this kind of money into mega-project ventures, you are bound to get some hits and misses. Just ask anyone investing into startups, because that's basically what they were doing but on a more grandiose scale. They bit off more than they could chew but the important thing is they have redefined their goals and targets (or scaled them back) into something that is easier to track, grow and scale relative to their abilities. Anyways, these articles are clearly written by people with no practical business experience because this happens all the time; leadership has grand visions, they shove money into projects, things don't quiet turn out as expected, you get pullbacks or outright cancellations and reposition. This is basically the learning curve, you are basically being asked to build out a whole technology/industry ecosystem in one go. There are bound to be flops, and they won't be the last. But the experience of the frauds and failures of yesterday will make the survivors stronger tomorrow.
Not even an expert or has his PhD yet because it’s ongoing.
It’s disappointing we will likely see his articles again. Turns out he has connections to the CSIS.
I already have the article prepared:
"Chinese might be able to make the semiconductors, probably using western recipes and IP, but the secret is in designing the chips. Every single one of the 10 billion transistors on a chip has a purpose. A single one malfunctioning can cause total failure. And of course even the best chip is useless without software. The chip design and software industries are far larger than the fabs, fitting their status as high end, high IP activities rather than dirty manufacturing.
I believe it will be centuries before Chinese can design and write software for the extremely complicated chips that require the entire world's accumulated IP to design."
Lol, I don't know why people constantly over-estimate Western political institutions like this. It's not some 4D chess move to convince people that China sucks at tech. They just genuinely don't know. Just think about how many terrible analysts and think tanks there are. Now how many of them speak and read Chinese?
There simply aren't enough experts in the field to give Washington any accurate information about Chinese efforts in semi-conductors.
This is what the "Western assessment" is in 2023. Prepare for a "Western assessment" in 2025-2030 to be all shocked about how quickly China caught up and replaced their dependence on Western components. And why this justifies an even bigger escalation to the trade war and tech embargoes. Except it will be way too late by then.
Well a pretty big problem with these analysts is that China is straight up hiding things. They discovered that China can produce 7nm chips by cutting open a chip SMIC made a year after it was in production.
Even what we are reading here, we need to take with a grain of salt. Ultimately, it's in China's best interest not to reveal too much.
In fact that's also the only reason I'm here. It's quite apparent that this forum gets some information from leaks like Havok which frankly is the best anyone can actually get at the moment.
I would not be surprised if Huawei's new chip set is a 3D layered 7nm chip which would put it pretty close to even the most modern chips in performance but not power consumption.
It may not be perfect but from my estimates, China can fully produce 28nm completely in house. Though yield is going to be questionable. But that said, since they kind of started from 0, that's pretty amazing.
- Status