Chinese semiconductor industry

Orthan said:

Im making this question because people here are making an assumption that huawei is going to be able to continue to produce its 5nm chips moving forward, without having a clue as to how.

Click to expand...

1. It's not an unreasonable assumption based on large trends. You don't need to be a top scientist with precise understanding of the lithography techniques to see the large trend that China overcomes every ban with its ingenuity. That is a large trend and everything now suggests that it will continue.

2. 5nm eventually, but I don't think anyone is talking about that coming into play very soon. People are mostly talking about SMEE's 28nm DUV and EUV projects set to complete in 2021 and 2022 which would be capable of making 7nm chips and likely 5nm chips going forward. So this isn't "without having a clue as to how" but rather piecing together the evidence to take an educated guess. On the other hand, you're the person who has no rational reason (other than that it's really hard) but believes based on sheer desire that China can't do world-leading lithography despite its current unprecedentedly aggressive push and record of overtaking global leaders in tech.

3. Your question is still stupid because it's not meant to be answerable with the public information available, at least not in greater detail than above. It takes a scientist with top clearance among China's national semiconductor projects to answer with a plan covering meaningful detail; such a person does not exist on this forum nor would he be inclined to divulge such information if he did. You asked this question because it's impossible to answer with what we have and you hope that that would make your opposite assertion look right... at least until a year or so later then it's proven wrong. It is like a caveman asking his fellow cavemen at night to explain the scientific mechanism by which the sun rises in the morning and if they can't, then they should believe his assertion that it will not rise tomorrow... until it does.

Orthan said:

Perhabs. Even so, TBH, at least i know who they are and i tend to give them more credibility than people in an internet forum that i dont know anything about, and perhabs are as clueless as i am.

Click to expand...

Fair enough.

Orthan said:

IMO, those examples dont work here because the chinese space/military industries doesnt really have to compete abroad. They exist to serve china´s needs. The semiconductor business is different. Even if china is able to replicate the complete chip production chain, they will need to be competitive, unless china bans/strongly discourages foreign chips. Of course, the chinese consumer could suffer from that, but thats another matter.

Click to expand...

It's highly relevant here. The reason that China had not taken the completely top-down approach is precisely because it wants the market to play the primary role, with the state playing a supporting role. Under this approach, China has been making rapid progress, particularly in the last five years. It starts to dominate the downstream industries (smartphones, consumer electronics, etc.) and it has made significant inroads in the middle-stream industries (SMIC, Hua Hong, Yangtze Memory, ChangXin, etc.). It starts to catch up with the world-class in IC design (e.g. HiSilicon). It's moving in the upper stream industry, semiconductor equipment and EDA, where it is relatively weak, but has players in every area. This would be a more organic approach of developing the entire semiconductor industry.

But the broad US sanctions have changed everything. It's not just Huawei now, but every Chinese semiconductor company is under threat. Now the goal is no longer always picking the vendors & partners with the current best technologies. The priority now is NOT for Huawei or SMIC for example to chase the latest and greatest processing nodes (although they would still pursue that goal to the extent they can), but to build self-sufficient, indigenous semiconductor supply chain. That's why we have been talking about China building indigenous fabs at 28nm/7nm and beyond in 3/5/10 years time frame in this forum recently. They would very likely still be behind the cutting edge then, but would be indigenous and sanction-free. The good thing, unlike Japan in the '80s when it got beaten by the US in the semiconductor industry, is that China has the world's largest semiconductor market to sustain such a strategy.

In other words, it's very much like the space/military industry, with the added benefit that it is still civilian and open to everyone else except the US-controlled interest. The best of the both worlds. That's why many Chinese companies are thanking Trump administration for its sanctions. Because they now have much better customer access and opportunities than they would do in a complete open and pure market competition environment.

Orthan said:

Im making this question because people here are making an assumption that huawei is going to be able to continue to produce its 5nm chips moving forward, without having a clue as to how.

Click to expand...

I don't think anyone here knows exactly what Huawei would do or would not do. But then again, a lot of people's confidence in Huawei's future as a tech giant is not blind. Huawei's premium smartphone business will undoubtedly be affected by the availability of the cutting edge chips, but its most strategic business, telecom equipment & 5G etc, will not be affected much by the sanction. For one, Huawei's smartphone business needs hundreds of million of chips a year and the chips have to updated every year, while Huawei only needs to stockpile millions of chips for its 5G business over the life of 5G deployment in the world, not every year. And, 5G base station does not need the latest semiconductor processing node, so they can be made by Chinese fabs or Huawei's own fab in a few years.

Orthan said:

Im making this question because people here are making an assumption that huawei is going to be able to continue to produce its 5nm chips moving forward, without having a clue as to how.

Click to expand...

This how it works in business and in economics.

You are suppose to not know, then it happens.

Where did Google or the internet come from? It came from nowhere, then it came, and it exploded outwards, and now it is everywhere.

But there were clues that this could happen, the research behind the scenes.

So when Google and the internet came out, that was totally unexpected, yet it was not a surprise.

It was not a surprise because of all the activities behind the scenes.

I think we have been talking about this all this time?

My best guess, it should be a timeline, and I estimate 1 year to 2 years maximum.

Others may disagree, such as Liberal media, whose timeline for this happening is forever, and America will be number #1 forever. This is the same America who cannot build a standalone 5G network under any circumstances, and manufactures exactly zero cell phones!

Then when that is pointed out, the voices in the Liberal media start with the excuses. The same excuses that give America the advantage, but give China the disadvantage (even though the excuses are the same).

After the excuses ... when it is not working out ... we know what comes next ...

Attack by cheap political attacks!
Tariffs!
Disinformation campaign!
Sanctions!

What else is new? CIA is always CIA.

manqiangrexue said:

People are mostly talking about SMEE's 28nm DUV and EUV projects set to complete in 2021 and 2022 which would be capable of making 7nm chips and likely 5nm chips going forward.

Click to expand...

Well, i could be wrong in this but can you provide a link to the information that a chinese made chinese EUV is going to be completed in 2022?

manqiangrexue said:

Your question is still stupid because it's not meant to be answerable with the public information available

Click to expand...

Just because no one has a clear answer for that question in this forum means that its a stupid question? How old are you? You need to take a deep breath and learn more about life in general, you know.

weig2000 said:

They would very likely still be behind the cutting edge then, but would be indigenous and sanction-free. The good thing, unlike Japan in the '80s when it got beaten by the US in the semiconductor industry, is that China has the world's largest semiconductor market to sustain such a strategy.

Click to expand...

It seems to me that you are repeting what i said, and ignoring again that unlike the space/military industries, the chinese semiconductor industry will have to compete internationally. If it is behind the cutting edge like you said, and unless foreign chips are banned/strongly taxed (which by itself raises questions about the viability of getting foreign markets open), then why would chinese consumers choose inferior chinese-made chips? For patriotic reasons, perhabs.

horse said:

You are suppose to not know, then it happens.

Click to expand...

That made me laugh. Keep posting stuff like this. It keeps things funny in this forum.

Orthan said:

It seems to me that you are repeting what i said, and ignoring again that unlike the space/military industries, the chinese semiconductor industry will have to compete internationally. If it is behind the cutting edge like you said, and unless foreign chips are banned/strongly taxed (which by itself raises questions about the viability of getting foreign markets open), then why would chinese consumers choose inferior chinese-made chips? For patriotic reasons, perhabs.

Click to expand...

It appeared that we just missed by a mile. Sigh. I now feel it's been a huge waste of time.

I'll rest my case with you on this topic.

Orthan said:

That made me laugh. Keep posting stuff like this. It keeps things funny in this forum.

Click to expand...

That's reality. And you cannot accept it.

New things come out of nowhere all the time.

In your world, does the "Eureka" moment exist?

Obviously not!

Semiconductors Are China’s Choke Point
By
October 24, 2020, 6:45 AM EDT

The rule of thumb in the semiconductor industry—Moore’s law—states that the number of transistors on a microchip doubles every couple of years.

That figure, improbably, . Meanwhile, the tally of companies that can afford to keep pace with the investments necessary to pile all those transistors on to a silicon wafer has moved just as dramatically in the opposite direction: only three remain at the cutting edge of manufacturing, down from about 25 in 2002. .
The industrial future of the planet rests to an extraordinary degree on two companies— and Taiwan Semiconductor Manufacturing Co., the go-to supplier for Apple Inc. and the focus of next-generation chipmaking. Both rely on U.S. technology.
This is China’s economic choke point. Beijing needs the most advanced chips to unlock the , artificial intelligence and quantum computing, but its attempts to build them are running up against the hard constraints of finance, physics and geopolitics.

Finance is the easiest challenge: the Chinese government has already earmarked more than $200 billion for the effort—a sum that exceeds, in real money terms, what America spent on the Apollo moonshot. The latest semiconductor factories, or fabs, will soon cost about $16 billion.
The physics are far more complicated. “Despite massive government investment in semiconductors, it is unlikely that Chinese companies will enter the top tier of global semiconductor manufacturers over the next decade,” according to a . For the foreseeable future, China will be at the mercy of America to remain globally competitive.
That’s where geopolitics come into play. Semiconductors are at the heart of the U.S.-China security competition; they drive industrial robots as well as rockets and submarines.

Two questions now dominate the outlook for global technology. First, how hard does the U.S. press on China’s semiconductor choke point, and second, how strenuously China resists. and its subsidiaries already threaten to cripple China’s technology leader. And although Beijing has yet to retaliate against U.S. tech companies in China—or choke off the supply of rare earths needed to manufacture chips— Beijing’s restraint .
Nor can its patience with Taiwan. The colossal imbalance in the semiconductor industry between the exponentially increasing power of chips on the one hand, and the concentration of manufacturing on the other, has heightened tensions in one of the world’s last remaining Cold War-era flashpoints.
Taiwan Semiconductor has become “turf that all geopolitical players ,” the company’s founder, Morris Chang, said in November.

A Chinese invasion to grab and occupy that territory is unlikely, though not altogether unthinkable. China has never ruled out the use of force to unify with Taiwan, which it still views as a renegade province. And Chinese warplanes have been intruding into its airspace in recent weeks as the administration of U.S. President Donald Trump draws closer to Taipei.
The more likely outcome, however, is a technological decoupling that would divide the world into “blue” (American) and “red” (Chinese) supply chains, weakening the R&D capabilities of U.S. chip companies by cutting them off from their biggest clients in China, raising costs for businesses everywhere .
Moore’s Law remains on the books. But its promise of exponentially expanding technological advance is hitting a wall.

From Beidou 2020. Still small but China does have domestic supplier in each category of semiconductor process

In 2019, there were $13.5 billion in imports of foreign equipment, but home-grown equipment sales were less than $250 million from leading companies AMEC and NAURA, and others including ACM Research, Mattson, and Shenyang Piotech, according to our marketing report. The next table shows the various types of equipment manufactured by Chinese suppliers and comparing them to the top foreign supplier and that supplier's market share.

Table 1: 2019 Market Shares: Chinese vs. ROW Semiconductor Equipment Revenues



Note that (1) there are few Chinese companies for each type of equipment and (2) the top foreign equipment supplier has significant market shares, a deep moat for Chinese suppliers.

As for customers:

• AMEC’s etch system is used in TSMC’s 5nm fab and is developing a high aspect ratio etcher and staircase etcher for 128-layer 3D NAND manufacturing at YMTC. Other customers include SMIC, Huahong, and Huali.
• NAURA has a large product offering, and its customers consist of SMIC, Hua Hong, YMTC, and GTA Semiconductors.
• Whereas NAURA sold 8 etch systems and 6 CVD and ALD deposition systems to Chinese semiconductor companies, the company sold 34 furnaces in 2019 as well as 16 cleaning systems.
• Shenyang Piotech received orders for 4 PECVD (for SiN, SiO2) systems from YMTC, and also is receiving repeat orders from Hua Hong and SMIC.
• ACMR competes in the cleaning system sector. They have been installed by YMTC and Hua Hong/Huali as well as SK Hynix (
  Please, Log in or Register to view URLs content!
  ).

NAURA is capable of producing chips at 5nm, comparable to those of AMAT and peers. NAURA is making equipment with 14nm capabilities but is developing etchers and deposition equipment for 7nm and 5nm nodes. It’s important to recognize three things:

• Please, Log in or Register to view URLs content!
  . For example, Intel’s 10nm is slightly denser than TSMC’s 7nm for SRAM. But TSMC’s 7nm is actually denser than Intel for logic.
• Although Chinese suppliers have 5nm capability, it is doubtful whether the 7nm node can be reached without EUV,
• Just 25% of China’s chip capacity is <20nm, technology about six years old

The previous table shows that in Q1 2020, just 1.3% of SMIC’s revenues were from chips made at 14nm.

Many Chinese chip equipment and materials makers have also gone public, and received massive financial support from the local capital market. The future may be bright, even though the scale of these companies is still small compared with market leaders. SMIC, the leading Chinese chip supplier is incented by the government to work with them and test local suppliers, but we don’t expect these companies to replace any of the leading suppliers soon. The next figure sums up the huge lead foreign equipment have over native Chinese companies. In 2019, Chinese suppliers sold just $200 million worth of equipment compared to imported equipment from foreign suppliers valued at $13.3 billion.

Figure 2: Semiconductor Equipment Revenues 2019 China vs. ROW




Although Chinese-made equipment has the ability to be used in the fabrication of state-of-the-art 5nm chips, there are numerous other factors that semiconductor manufacturers use to evaluate a supplier:

• Reliability
• Uptime
• Price/performance ratio
• Mean time between failures (MTBF)
• Equipment support
• Limited breadth of equipment offerings within a sector

But the Chinese government is preparing broad support for so-called third-generation semiconductors for the five years through 2025.

Semiconductor manufacturers typically take 9-12 months to evaluate a piece of equipment and make decisions on a "best-of-breed basis." However, the Chinese government is demanding that a portion of equipment used in a fab must be Chinese made. If Chinese made equipment is sitting side-by-side to a foreign-made system, how difficult would it be for Chinese equipment engineers to keep their eyes and ears open to discover features of the foreign equipment that could be implemented into their own.

If foreign engineers were not permitted by the Commerce Department to monitor equipment as occurred in October of 2018, when the Commerce effectively shut down Chinese semiconductor firm Fujian Jinhu by cutting off U.S. suppliers, including suppliers of semiconductor-making machines because the firm allegedly stole U.S. memory chip maker Micron’s (NASDAQ:

Please, Log in or Register to view URLs content!

) technology, foreign equipment engineers picked up their tools and exited the building. Questions were raised as to what happened to that equipment and did the company reverse engineer the equipment already in the fab. Applied Materials would be most impacted by export restrictions of all foreign suppliers for several reasons:

• The company supplies more types of equipment to China and has a larger installed base of equipment than any foreign peer. AMAT makes every type of equipment except for lithography and resist processing.
• AMAT makes more variations of equipment within a sector than peers. For example, in the CVD sector, AMAT offers 12 different types of equipment compared to just 5 for LRCX and 4 for TEL. In the PVD (sputtering) sector, AMAT has 16 different variations of equipment while China's NAURA has just 6.

Commerce restrictions could halt AMAT's acquisition of Hitachi Kokusai Electric as Chinese regulators still need to approve the acquisition and so far, they have refused.

The biggest obstacle for Chinese chip company is continuous availability of foreign chips. Chip by nature is a nature monopoly, device makers naturally will go for the best one in the market, as the power efficiency and performance of chip generally supersedes the cost of chip, considering the total device cost is much higher than the chip within. Trump has no bloody idea of how the industry works. What Chinese chip industry needs is market and time, now they have both.





China's orbit contains the majority of foundry activity. This is advantageous to China.