Chinese semiconductor industry

Nutrient said:

I think I understand what you mean now. My impression is that China didn't focus on even core technologies because it could not do so: 15 years ago, the underlying technological base did not exist. So the Party developed what it could, the more software-like chip designs. I would not call it a policy failure; it was rather a lack of feasible alternatives.

Now the situation is different. The Middle Kingdom has comprehensive technologies, or will soon develop them with internal resources. Total self-sufficiency is possible in semiconductors now.

Click to expand...

Imo, I absolutely consider it a policy failure because it is the China government relying on the chicken and egg situation while ignoring the strategic realities. In a geostrategic situation like we've had since the Pivot to Asia, there is literally no choice but to invest heavily even in those things that seem impossible. The alternative is surrender. Since China didn't bow down before America like Japan did, they should have naturally went all out. Instead, China continued to underfund core technologies in order to maximize profits in the most feasible commercial areas it was closest to achieving. This is exactly what American strategists count on to happen. I'm not advocating ONLY investment in core technologies and ONLY focusing on weaknesses. This is not a binary choice. However, China made the fatal mistake of trusting globalization as if the American control freaks would never touch it because they've been promoting it since WWII.

Those previously believed impossible goals of indigenizing semiconductor equipment and materials was once believed to be out of reach for at least 10+ years, if ever. That was the consensus opinion from industry experts only 3 years ago! It wasn't until Trump used the sledgehammer that it forced China on the path it is now. If industrial policy emphasized semiconductor equipment and materials back in the early days of the Pivot to Asia, we would be 5 years earlier in the game and China would have DUV comparable to Nikon/ASML, TSMC/Samsung now and probably have commercialized EUV by 2020.

krautmeister said:

Imo, I absolutely consider it a policy failure because it is the China government relying on the chicken and egg situation while ignoring the strategic realities. In a geostrategic situation like we've had since the Pivot to Asia, there is literally no choice but to invest heavily even in those things that seem impossible.

Click to expand...

Note that Grace Semiconductor (Shanghai) was founded in 2003, years before Obama's election and the Pivot to Asia. So the CPC foresaw the new strategic realities. The company was not successful because China simply did not have the necessary technological base back then. The inability to localize semiconductor manufacturing was not a policy failure at all; the government clearly wanted to do it but could not.

Nutrient said:

Note that Grace Semiconductor (Shanghai) was founded in 2003, years before Obama's election and the Pivot to Asia. So the CPC foresaw the strategic realities. The new company was not successful because China simply did not have the necessary technological base back then. The inability to localize semiconductor manufacturing was not a policy failure at all; the government clearly wanted to do it but could not.

Click to expand...

@Nutrient bro is Hua Hong an offshoot from Grace Semiconductor?

ansy1968 said:

@Nutrient bro is Hua Hong an offshoot from Grace Semiconductor?

Click to expand...

That could be. A company named Shanghai Huahong Grace Semiconductor Manufacturing exists; it probably has some connection to Grace.

Nutrient said:

That could be. A company named Shanghai Huahong Grace Semiconductor Manufacturing exists; it probably has some connection to Grace.

Click to expand...

@Nutrient bro I think Grace Semi had serve its purpose aside from Hua Hong there is Huali and others. So my apologies to Jiang Zemin for calling him a buffoon, he and his son did their best with limited tools at their disposal.

jfcarli said:

This is not entirely bad news. YMTC is also testing 192 layers which is more than SAMSUNG can presently do. If they manage to sort out yields of both 128 and 198 layers, YMTC will become the top dog as far as NAND memory is concerned.

Please, Log in or Register to view URLs content!

Click to expand...

I thought Samsung (980 pro) and Micron already use 192 layers? What brands are using NAND from YMTC

So far I only used Intel 5-6 yrs ago and then only Samsung, the latest one I bought 1 TB and 500GB Samsung 980 Pro for my newly built desktop (by myself) with Ryzen 5900X 12 cores/24 threads

Interesting 4 page report on SMEE and 28nm lithography equipment.

What is particularly worrying is:

"
1. Optical systems

A representative of Changchun UP Optotech Co., Ltd., SMEE's optical system supplier, (link in Chinese) that the company did not participate in the latest lithography machine research project. Another supplier, Beijing GuoWang Optics Co., Ltd., (link in Chinese) that it is expecting to finish its foundry construction by September 2023. This new plant will make possible the mass-production of the optical system for ArF immersion lithography. With its key suppliers in this segment facing some difficulties, SMEE might not complete a new ArF immersion stepper this year.

2. Light sources

SMEE's primary light source system supplier Beijing RSLaser Opto-Electronics Technology Co., Ltd. (RSLaser) has delivered the first domestic 40W 4kHz ArF light source. However, TWINSCAN NXT:1980Di, an ASML-designed machine currently used to make chips at 28nm, employs a 60W 6kHz ArF laser. Now, RSLaser's team has to make a giant technical leap, so that SMEE can finish the whole machine on time."

Here is the report

antiterror13 said:

I thought Samsung (980 pro) and Micron already use 192 layers? What brands are using NAND from YMTC

Click to expand...

I wouldn`t bet my house on this, but I am pretty sure Samsung and Micron most advanced 3D NAND chips are 128 layers. Seems that both Samsung and Micron are developing 176 layers device.

YMTC is currently developing 192 layers.

It seems, however, that YMTC`s yield in 128 layers is quite low and not comparable to either Samsung or Micron.

Considering however that YMTC was founded in 2016 and is fully producing 64 layers, I am pretty sure it won`t take long before their 128 layers chips will be on par with the current incumbents.

YMTC has its own SSD brand Zhitai and I think Huawei already uses YMTC`s 64 layers in the Mate 40s.

Nutrient said:

Note that Grace Semiconductor (Shanghai) was founded in 2003, years before Obama's election and the Pivot to Asia. So the CPC foresaw the new strategic realities. The company was not successful because China simply did not have the necessary technological base back then. The inability to localize semiconductor manufacturing was not a policy failure at all; the government clearly wanted to do it but could not.

Click to expand...

You're talking about the early 2000s before China's overall scientific-technological base achieved critical mass. It was only sometime between 2005-2015 when China achieved critical mass in almost every field. I remember the fanfare given to Grace Semiconductor when it was founded. Jiang Mianheng was a co-founder and the government was giving it full support and providing all the funds it needed. Soon afterwards, SMIC was also founded. Despite the government obviously favoring Grace, SMIC quickly surpassed it technology-wise and has remained at China's leading edge ever since. So, having the required skills and experience is definitely required. The Pivot to Asia began in 2012 and its beginnings were already brewing in 2009. China already had the ability for some catchup by around that time and if it began at that time, it would be several years ahead of where it is now, there's doubt about that. China's policy failure was that it took so long to realize their vulnerable position because the relevant decision makers believed so strongly in globalization that it blinded them to reality.

I compare this to China's family planning policy makers. They kept the 1 child policy in place all the way until 2015 even though it was blindingly obvious that a demographic cliff was already under way 7-10 years before. Then they expanded it to 2 children and it barely made a dent. It took a full 6 years to jump from a 1 to 3 child limit. I consider this negligence.

jfcarli said:

I wouldn`t bet my house on this, but I am pretty sure Samsung and Micron most advanced 3D NAND chips are 128 layers. Seems that both Samsung and Micron are developing 176 layers device.

YMTC is currently developing 192 layers.

It seems, however, that YMTC`s yield in 128 layers is quite low and not comparable to either Samsung or Micron.

Considering however that YMTC was founded in 2016 and is fully producing 64 layers, I am pretty sure it won`t take long before their 128 layers chips will be on par with the current incumbents.

YMTC has its own SSD brand Zhitai and I think Huawei already uses YMTC`s 64 layers in the Mate 40s.

Click to expand...

How about the quality of YMTC NAND compared to the top dog Samsung ?