Chinese semiconductor industry

BlackWindMnt said:

Why would China play ball when the EU is so hostile, it would be better to start JV with Russians to say like the 28nm level it will boost Russian capabilities while not giving away the high end marketshare they will probably erode from TSMC, Samsung and Intel.

Click to expand...

@BlackWindMnt brother for me a sweet victory is an all Chinese FAB in EU. Imagine a SMIC FAB in Poland using All Chinese made equipment producing 28nm , 14nm chips or even 7nm chips Operate mainly by Chinese, you can imagine the uproar it will create....lol

EU had an opportunity to be the third bloc, Xi had convey this message to Merkel last year after signing CAI, BUT She choose the easy way out and act in America interest instead of EU.

AndrewS said:

It's not just a matter of cost to find a Huawei alternative.

In many places, you're talking about a 2+ year delay in the deployment of standalone 5G networks in Europe.

If European telecoms networks had been able to use Huawei, European companies potentially could have edged out US companies in the development of 5G applications, because the US telecoms industry had hobbled itself, for a variety of reasons.

Those 2+ years mean US companies are more likely to deploy 5G applications at scale first, and then colonise the European market like what happened with 4G and the development of smartphone apps.

Click to expand...

Good post. When Europe can't understand the simple matter of 5G and the consequences of banning Huawei, what is the hope that they will somehow keep their word on a potential semiconductor JV.

Europe due to its partnership with Huawei leapfrogged the US and was in prime position (along with China) to dominate 5G next gen industries/applications.

Instead, the very neutral and totally not a US puppet EU, threw all that away in order to harp about "Huawei is spying us" "Evil CCP is everywhere".
What they accomplished with their lies and propaganda is that China will dominate 5G, followed by the US, while the EU gets technologically colonised by the US once again(see Big Tech companies...)

For people who want to learn more I would recommend periodically checking the World News thread in order to keep up to date with the geopolitics today.

To conclude, SMIC building a fab in the EU is fantasy. Maybe after 2030/35 it could happen, depending on geopolitics (not on economics)

ansy1968 said:

@BlackWindMnt brother for me a sweet victory is an all Chinese FAB in EU. Imagine a SMIC FAB in Poland using All Chinese made equipment producing 28nm , 14nm chips or even 7nm chips Operate mainly by Chinese, you can imagine the uproar it will create....lol

EU had an opportunity to be the third bloc, Xi had convey this message to Merkel last year after signing CAI, BUT She choose the easy way out and act in America interest instead of EU.

Click to expand...

I don't get it at all. How could an all Chinese FAB in EU be a sweet victory? Why would SMIC have a FAB in Poland? Why would SMIC want to send such sensitive equipment and important people to EU only to make them susceptable and vulnerable to robbery/abduction by the EU/US? What does SMIC gain from doing such a thing?

bettydice said:

I don't get it at all. How could an all Chinese FAB in EU be a sweet victory? Why would SMIC have a FAB in Poland? Why would SMIC want to send such sensitive equipment and important people to EU only to make them susceptable and vulnerable to robbery/abduction by the EU/US? What does SMIC gain from doing such a thing?

Click to expand...

@bettydice Sorry bro it was a hypothetical answer from a hypothetical question by @european_guy, so my apologies for deviating and now back to normal programming.

Primarius NanoSpice™ Simulator to Accelerate the verification of CMOS Image Sensor is adopted by Skhynix​

Primarius Technologies today announced that its high-performance parallel SPICE simulator has been adopted by SKhynix to accelerate the verification of CIS chips. NanoSpice™ helps SKhynix’s customers to verify their designs with efficient CPUs arrangement, accuracy confidence, and better convergence on default mode.

CIS chips with the capacity and significantly improved level of quality, require fast and accurate circuit simulation and verification tools to support advanced IC designs. provides good simulation convergence and accuracy, which helps SKhynix’s customers to design with confidence and shorten design cycle.

As a new generation parallel SPICE simulator, was designed for the most challenging simulation with high-capacity and high-performance, such as large post-layout analog circuit simulations requiring capacity, speed, and accuracy simultaneously.’s superior parallelization technologies handle generic circuit simulation with up to 50M+ circuit elements.

“SKhynix and Primarius have achieved joint success with , which has embodied mainly in three aspects. is so efficient that could be well used with 16 cores from just 1 license. And without any conditions, its default is as accurate as reference. Most importantly, it is usefully verified with CIS verification part.said Mr Changho,Do CAE VP at SK hynix.’

“SK hynix has been a long-time customer of our EDA products and we continue to improve our technology to meet the simulation challenges presented by CIS and Memory chips.” remarks Dr. Zhihong Liu, Primarius’ Chairman and CEO. “ fits the performance and accuracy requirements of the toughest design verification needs with its full range of simulation capacities from small blocks to full chip designs, and can help achieve goals that would be difficult to reach without our circuit simulation solutions.

Shengmei Shanghai received a bulk purchase order for Ultra C wb trough wet cleaning equipment​

Shengmei Semiconductor Equipment (Shanghai) Co., Ltd. (hereinafter referred to as Shengmei Shanghai) (Science and Technology Innovation Board stock code: 688082), a leader in providing wafer process solutions for semiconductor front-end and advanced wafer-level packaging (WLP) applications The supplier, announced today that it has received a bulk purchase order for 29 Ultra C wb tank wet cleaning equipment, which can be used to process 300mm wafers, of which 16 repeated orders for the equipment are from the same domestic foundry in China , the purpose of the repeat order is to support the expansion of the plant. The devices are scheduled to ship in two phases starting in 2022.

Dr. Wang Hui, Chairman of Shengmei Shanghai, said: "We developed the trough wet cleaning system to cover the needs of customers for wet cleaning technology under a wider range of technology nodes. This order proves our highly competitive products. , the strategy of making the company's product portfolio complete from advanced cleaning process to full cleaning process coverage is correct. Obtaining this repeat batch order from the customer proves that our strategy is successful - because this order is from the same foundry, And it belongs to the wet trough cleaning market category. This also clearly demonstrates our technical strength, growing market leadership and ability to meet customer needs. Coupled with Shengmei Shanghai's advanced low pressure drying technology, our trough products are almost All cleaning process steps using tank cleaning equipment can be completed.”

ULD (Ultra Low Pressure Drying) technology

The advanced drying technology mentioned by Chairman Wang Hui above is the ULD (Ultra Low Pressure Drying) technology for 300mm trough system that Shengmei Shanghai announced today. This process is specially designed to solve the technical challenges of drying in bath cleaning, such as drying of 3D NAND structures on advanced semiconductor wafers and high aspect ratio structures of logic products in bath cleaning. Shengmei Shanghai's new ULD tank module uses a low pressure isopropyl alcohol (IPA) drying process to meet the requirements of most tank cleaning and drying processes, including pre-furnace tube cleaning, post-ion implant cleaning and post-dry photoresist Removal, post-chemical mechanical polishing (CMP) cleaning and processes such as thin film deposition, oxide etch and nitride removal. Shengmei Shanghai has delivered its first ULD module to a leading Chinese memory semiconductor manufacturer in the third quarter of 2021, and preliminary process data has demonstrated the effectiveness of ULD in advanced node manufacturing.

Agree China should focus on commercializing lagging nodes like 28nm first, there is significant demand in various Power ICs/RFs semis, don't require the most advanced equipment, with the narrowest of tech gaps. It's not just about having the technology, but also doing it at commercially viable production yields. You can subsidize for a while, but it's unsustainable in the long-run. For example, DUV can theoretically handle 7nm, but it would require 3 to 4 layerings (vs. 1 layer for EUVs) and be completely unviable commercially. Imagine being 4x slower to produce the same product.

Military/government applications can rely entirely on reliability, but rest of the private sector needs to consider commercial angles. It's the same reason why WS-20 already flying for Y-20 strategic transport, but China is still relying on LEAP engines for C919 and potential C929 - military planes don't need to worry about cost per passenger-seat-mile.

SinoaTerrenum said:

Agree China should focus on commercializing lagging nodes like 28nm first, there is significant demand in various Power ICs/RFs semis, don't require the most advanced equipment, with the narrowest of tech gaps. It's not just about having the technology, but also doing it at commercially viable production yields. You can subsidize for a while, but it's unsustainable in the long-run. For example, DUV can theoretically handle 7nm, but it would require 3 to 4 layerings (vs. 1 layer for EUVs) and be completely unviable commercially. Imagine being 4x slower to produce the same product.

Military/government applications can rely entirely on reliability, but rest of the private sector needs to consider commercial angles. It's the same reason why WS-20 already flying for Y-20 strategic transport, but China is still relying on LEAP engines for C919 and potential C929 - military planes don't need to worry about cost per passenger-seat-mile.

Click to expand...

I think for now Chinese companies should strive to use 100% indigenous SMEs and software on the older 350nm-45nm nodes and use as much local equipment as possible on the newer nodes. The point is that they need to develop a local ecosystem that can help them offset any future disruptions to their technology supply chain caused by overzealous politicians. Once the ecosystem is created, developing from there will be much easier.

european_guy said:

I agree that a JV endorsed by European Commision is a must, otherwise the project is doomed to fail.

Click to expand...

History has shown that an endorsement from European Commision does not mean much. In 2003 the said comission signed an agreement with China on Galileo project. 3 years later after China has invested time and resources they were kicked out of it for "National Security" reasons.

SMIC is not the only one currently banned from owning ASML EUV. Even the foreign companies operating in China are banned from bringing it to their Chinese mainland fabs. Very recently a Chinese company tried to purchase the Newport Wafer Fab in UK. That fab's best node is 180nm and is running on loss. Wingtech was actually doing a favor on the fab and Newport economy by investing in it. However there was a major national uproar in UK because "china could steal the secrets of this 180nm fab", and it is still under National security review and the UK government could still take it back under a new law that was passed few years ago, even though it has already been paid for.

It is highly unlikely the US would allow a Chinese company to operate ASML EUV tools even outside the boarders of china. And if they used SMEE tools they can always say that the chips coming out of that fab are "backdoored", "national security", etc, etc.

SinoaTerrenum said:

Agree China should focus on commercializing lagging nodes like 28nm first, there is significant demand in various Power ICs/RFs semis, don't require the most advanced equipment, with the narrowest of tech gaps.

Click to expand...

Sorry to send another post on the same topic, I won't insist further.

I agree Chinese equipment manufacturers should focus on 28nm, that would already be a huge success. The last one to have all the supply chain in house was US in the '70-'80, so it is at least 40 years that nobody is able to do that.

But I'd just would like to highlight that there is a whole emerging semiconductor design industry in China that currently (not in the future) already requires better nodes. For instance China is at the forefront of AI and especially AI devoted to autonomous driving. China is at the forefront of cloud computing, China is at the forefront of smartphone's base-band processing (CPU + MODEM chip).

If for some reason someone (the usual one) decides that TSMC and Samsung cannot produce 7nm nodes or better for Chinese customers, this means that China suddenly, from one day to the other, is cut out from the most important technology of our near-mid future. This would be a devastating blow, considering that in these technology, like autonomous driving and EV vehicles, China is currently already leader or among the leaders. This is not a remote possibility, US is already actively working on it:

Please note, these are not fields in which China is catching up like in semiconductor manufacturing. These are fields in which China is at the forefront now (although at the design level, not at manufacturer level). To be stopped for 5 years it means to start again from behind and loose all the first-mover big advantage that China has today.