It is reported online that Zhang Rujing said that "the United States has little power to restrict China", and the organizer responded
[Copy link]
On August 8, Jinshajiang Capital issued a statement by CITIC Securities and Jinshajiang Capital on media reports of Zhang Rujing’s remarks. The following is the full text of the statement:
Recently, the media published articles with titles such as "SMIC founder Zhang Rujing shockingly said: The United States' ability to restrict China is not that strong". The articles attracted attention with the titles, but the titles did not correspond to the focus of the conference. We understand the good intentions of the author, but it may have some adverse effects on the development of third-generation semiconductors and the commentators. After detailed communication with the guests and friends in the industry, we decided to publish an article to express our opinions; the commentators accepted the invitation of CITIC Securities and Jinshajiang Capital to participate in the third-generation semiconductor online conference. The original intention was to share their views on third-generation semiconductors through the explanation of the knowledge of third-generation semiconductors, in order to help friends in the industry deepen their understanding and support for third-generation semiconductors. However, some media articles have too strong titles, one-sidedly emphasize the confrontation between China and the United States, and deviate from the theme and purpose of the conference. The development of semiconductors has greatly changed and enriched human life, relying on the cooperation of the global industrial chain. The recent Sino-US relations have broken this balance, but this abnormal confrontation may soon move towards rational cooperation. While China's semiconductor industry is relying on its own efforts and striving for "internal circulation", it should not forget to look at the world, learn from each other's strengths and weaknesses, and connect with friends and forces that can be connected.
The development of China's semiconductor industry depends not only on the efforts of semiconductor professionals, but also on the support of society and even the media. Whether it is about Sino-US relations or the development of China's semiconductor industry, we hope that everyone will treat it rationally and make rational analysis. Promoting long-term international cooperation and turning enemies into friends as much as possible is the right way to achieve mutual benefit!
So, what did SMIC founder Richard Chang say that caused the organizer to issue an urgent statement? Let's read on:
"If China maintains its lead in 5G technology, it will be far ahead in communications, artificial intelligence, cloud services, etc. in the future, because China is very strong in high-tech applications."
"If the United States cannot win in fair competition, it will take administrative measures, as it did to Japan in the 1980s. Starting in 2018, it began to restrict 5G again, but this time its opponent is no longer Japan. The United States' ability to restrict China is not that strong, but we cannot take it lightly."
"The third generation of semiconductors has a characteristic. It is not Moore's Law, but post-Moore's Law. Its line width is not very small, and the equipment is not particularly expensive. However, its materials are not easy to make, and the design must have advantages."
“For third-generation semiconductors, IDM is now the mainstream, but foundries still have opportunities. However, design companies need to find a foundry with which they can cooperate for a long time.”
"China is very strong in some areas, such as packaging and testing. As for equipment, such as lithography machines, we are far behind. If we look specifically at the materials, manufacturing, and design of third-generation semiconductors, I personally think that our gap in materials is not that big."
"We need to consider the talent base in a short period of time. This is one of our weaknesses. We may have laid the foundation, but there is a gap between the foundation and application. How can we shorten it? European and American companies do better, so we can learn from their strengths."
"Personally, I think that in the third stage, if we only have the materials and epitaxial wafers to make the device, if we use a 6-inch wafer, the investment depends on how big you want to make it. It will be at least more than 2 billion. You can make money if you reach a scale of 60 to 70 billion. This is the third stage. If the second stage is to make Epi (epitaxial wafers), the investment is not large either. The corresponding Epi factory investment will probably be less than 1 billion."
The above are the latest and exciting views shared by Zhang Rujing, the former founder and CEO of SMIC, general manager of Shanghai Xinsheng, and current founder and chairman of Xin'en (Qingdao), in the online live broadcast of the Wind 3C Conference recently hosted by CITIC Construction Investment Securities and GSR Capital.
In this rare sharing, he introduced the first to third generation semiconductors. During the question-and-answer session, Zhang Rujing combined the current background between China and the United States to share his views on domestic substitution of semiconductors, possible development models of third-generation semiconductors, and the gaps and advantages between China and foreign countries in various semiconductor sub-industries.
First to third generation semiconductors
Let me discuss with you some basic knowledge about the first to third generation semiconductors.
The first generation of semiconductor materials used was germanium, which was later changed to silicon. Because silicon has a large output and the technology is well developed, it has basically completely replaced germanium.
But below 40 nanometers, the application of germanium reappeared, and the germanium silicon channel can make electrons flow very quickly.
Therefore, silicon is the most commonly used material in the first generation of semiconductors. The germanium silicon currently used will be used in special channel materials, and in the future, carbon will be used. This is a later story.
These materials are all tetravalent in the periodic table, and group IV A carbon, silicon, germanium, tin, and lead all belong to the first generation.
The second generation uses compound semiconductor materials. We often use materials such as gallium arsenide or indium phosphide, which can be used in the field of power amplifiers. In the early days, they were relatively fast.
However, because arsenic is highly toxic, its use is now banned in many places, so the application of gallium arsenide is still limited to the field of high-speed power amplifiers.
Indium phosphide can be used to make light-emitting devices, such as LEDs.
By the third generation, better compound materials appeared, including silicon carbide, gallium nitride, aluminum nitride, etc., which are all formed by elements of groups 3 and 5.
Silicon carbide has special advantages in high voltage, high power and other fields; the switching frequency of gallium nitride can be very high, so it is often used in the field of high-frequency power amplifier devices; aluminum nitride is used in special fields and is less involved in civilian use.
There are also some special semiconductors that are not sure which generation they should be classified into, but in terms of time they span the second and third generations.
It is mainly concentrated in the elements of groups 2 and 6, and is formed by the combination of elements ranked II B and VI A in the periodic table, such as cadmium antimonide (CdTe), mercury antimonide (HgTe) or mercury cadmium telluride (HgCdTe), etc.
The uses of these materials are very special and the processing is relatively complicated. Basically, they are rarely used by the public, but they still fall into the category of semiconductors.
The above is a brief introduction to the third-generation semiconductor materials. We can try to discuss some details later.
Dialogue
The United States’ ability to restrict China is not that strong
All three stages of silicon carbide production may encounter bottlenecks
Q: Since ZTE was included in the US Entity List in 2018 and Huawei in 2019, the two hottest concepts in the semiconductor industry have been domestic substitution and overtaking on the curve. In the past two years, the entire Chinese semiconductor industry has faced blockades at all levels by the United States. What do you think of the so-called domestic substitution and overtaking on the curve in the Chinese semiconductor industry?
Zhang Rujing: I often don’t understand why we have to overtake on a curve. Can’t we overtake on a straight road?
In fact, you can overtake at any time. Overtaking on a curve is not a shortcut, but a time-consuming and laborious method. This is my opinion.
Currently, the United States has many restrictions on China's high-tech industry, but this did not start now.
As early as before 2000, the so-called Paris Coordinating Committee (abbreviated as "CoCom", which was disbanded in 1994) had already appeared. It was also an international technological blockade.
The most recent one is the Wassenaar Arrangement, which imposes an embargo on high-tech technology, materials and equipment on certain countries.
In 2000, when we returned to the mainland to build a foundry factory, these restrictions still existed. However, the Bush administration was relatively supportive of China and gradually relaxed some restrictions during his term.
At that time, if we at SMIC wanted to introduce 0.18-micron level technology, equipment and products to mainland China, we had to apply for a license.
We applied for 0.18 micron and 0.13 micron in the United States, and obtained the countersignatures of four departments of the U.S. government, including the U.S. Department of State, the Department of Commerce, the Department of Defense, and the Department of Energy.
The Department of Energy is special. It is afraid that we will make weapons such as atomic bombs, but we don't, so the Department of Energy usually approves us very quickly.
So this restriction has always existed, but it has been gradually reduced since 2000. We can apply for approval from 0.18 microns to 90 nanometers, 65 nanometers, and 45 nanometers.
Moreover, the 45-nanometer technology was transferred from IBM, which was quite advanced at the time. After that, we applied for 32-nanometer technology, which can be extended to 28-nanometer technology.
After that, I left SMIC, and I may not continue to apply for technology below 28 nanometers, but it may not be necessary.
In short, different American presidents will formulate different strategies for equipment restrictions, etc., and Trump’s strategy for China is the most stringent.
In the early days, the U.S. Department of Commerce was very supportive of us, and the Department of Defense had many opinions, but after discussion among four ministries, it was passed.
But the biggest resistance this time comes from the US Department of Commerce. This is mainly because the United States lags behind in 5G technology, so it hopes that China will slow down its development in the 5G field. There are precedents for such restrictions.
In the 1980s, we produced memory in the United States. At that time, Japan's memory technology was much more advanced than that of the United States in terms of technology, design, and yield. So the United States began to restrict Japan and later forced Japan to sign the Plaza Accord.
As a result, we can all see that Japan could not withstand the pressure from the United States. The memory industry almost died out, with only one or two companies still surviving.
In terms of logic, Japan has no special advantages and is not yet in the lead. Mainland China may be stronger than Japan in this regard.
However, Japan is leading in analog and mixed digital and analog technologies, and has done well in components for automobiles and high-speed rail.
At that time, under the influence of the United States, Japan was greatly restricted in all aspects and its development slowed down.
But Japan did not stop there, but instead devoted itself to the research of materials and equipment, so it is still very advanced in these areas.
For example, about 51% of the world's 300mm silicon wafers are produced by two Japanese companies, Shin-Etsu and Matsukouchi.
Japan is a leader in areas such as photoresists, specialty chemicals and materials.
In terms of equipment, Japan's lithography machines are also very advanced, and other basic equipment can also be produced independently, such as diffusion furnaces, LPC microcontrollers, etc.
But in general, after the United States imposed restrictions on Japanese technology, Japan's equipment, especially in the memory field, was hit hard.
However, this time when the United States discovered that China was putting great competitive pressure on it, the U.S. administrative leaders began to crack down on and restrict Chinese technologies, mainly 5G communications.
If China maintains its lead in 5G technology, it will be far ahead in communications, artificial intelligence, cloud services, etc. in the future, because China is very strong in high-tech applications.
Tik Tok, which has been making a lot of noise recently, is much better than Facebook in the United States. As a social networking site, Tik Tok has many interesting features and has suddenly become popular among many young people in the United States.
Of course, we can also see that the head of Facebook speaks very sourly, saying how bad China is, which is all out of jealousy.
If the United States cannot win through fair competition, it will resort to administrative means.
The United States did this to Japan in the 1980s, and began to restrict 5G again in 2018, but this time its opponent is no longer Japan. The United States' ability to restrict China is not that strong, but we cannot take it lightly.
The third-generation semiconductor materials are often used in the 5G field. For example, gallium nitride is used in 5G high-frequency chips.
This material has a very high frequency and can withstand high pressure and high temperature. For example, driverless cars or high-power charging piles will use silicon carbide, and the United States will impose embargoes on these materials to China.
Today, President Wu mentioned to me that the application of the third-generation semiconductor materials will focus on silicon carbide, so I said I would discuss silicon carbide with you more.
This is a very good material, but there are three stages in its production where bottlenecks can occur:
First, materials. Silicon carbide single crystals were used for a long time in the early days in sizes of 2 to 4 inches. Now, 6 to 8 inches have also been introduced. Of course, 4 inches and 6 inches are still the most commonly used. The raw materials themselves are very important resources.
The second stage is the production of epitaxial wafers, which also involves special technologies and is crucial to the quality of the final component products;
The third stage is to produce semiconductors of various powers, which have a wide range of uses.
The most widely used are high-power devices in new energy vehicles and trains, with voltages exceeding 3,000 volts, and are best made of silicon carbide. However, China's technology in this material manufacturing link is relatively weak.
What are the development laws of third-generation semiconductors?
IDM mode is now mainstream
Q: What do you think of the third-generation semiconductors? What rules will they follow in their development? Will the future development model of the third-generation semiconductors be dominated by IDM, or will the model of Design House (third-party design) plus Foundry (foundry) dominate?
Zhang Rujing: The semiconductor industry requires long-term investment, and practitioners must be able to endure loneliness. Experience is accumulated gradually, which is different from online e-commerce.
Those things can take off quickly if there is a good idea in a very short time, and investors are willing to invest money there.
However, the third-generation semiconductors have a characteristic: it is not Moore's Law, but post-Moore's Law. Its line width is not very small, and the equipment is not particularly expensive, but its materials are not easy to make, and it requires advantages in design.
The investment is not very large. So if it appears, first, is there a market? Yes; is there anyone willing to invest? Maybe some people are willing, because it is not a big investment and the rate of return seems good; is there government support? The government supports; is there a good team? This is a big question; are there enough talents? This is a question, there are not enough people with real experience in our country.
So for the third-generation semiconductors, taking silicon carbide as an example, the market for good products is very large because a lot of them are used in new energy vehicles.
For example, Tesla's Model 3 uses silicon carbide (silicon carbon) power module.
Who made these modules?
It is STMicroelectronics, and recently it has also started to buy some from Infineon. It is basically provided by these two companies, and these two are basically IDM companies, and it is doing very well.
It seems that the big ones in the third-generation semiconductors are all IDM companies, because they are responsible for the entire industrial chain from beginning to end, and their efficiency may be higher.
But there are also foundries. Some are in Japan, and some are good at Epi (epitaxial wafers). In China, we may not make the earliest single crystal substrates ourselves, but we make the epitaxial wafers on them ourselves.
After that, some design companies will design it and then have it taped out in different foundries. There are such silicon carbide foundries in Japan, Taiwan, and South Korea, so there is also this kind of division of labor and cooperation.
We also have people in mainland China who want to do this, so I personally think that for third-generation semiconductors, IDM is now the mainstream, but foundries still have opportunities. However, design companies need to find a foundry with which they can cooperate for a long time.
I personally think this is a good opportunity.
If the capital market is willing to invest, the capital required is many times lower than that required to build an advanced logic platform. It doesn’t take a lot of investment to do it, the key is talent.
We don't have enough of these talents in our country now, but we do have them in the United States. Some people are still willing to come to the mainland to work, but some in Japan may not be convenient for them to come.
There are such institutions in South Korea and Taiwan, and we also have some research institutes in mainland China that are doing quite well. If these people are willing to enter the industry, that would be great too.
I personally think that Samsung of South Korea has done a good job. Just now, several friends mentioned why Samsung has done so well.
He is an IDM with strong financial resources and a long-term vision. Although the domestic market is not large, he knows that he must master the technology. Therefore, Samsung China has long been developing materials and equipment, and has been involved in a variety of industries from start to finish.
If one day it is restricted and is not given this or that, basically it can survive except for the lithography machine.
Basically, it can produce first to third generation semiconductor products, and they are very competitive.
There is a company in Taiwan that is capable of doing what Samsung does. This company is very technologically advanced, but apart from technology, it basically does not develop much in terms of materials and equipment.
Because they think they will not be strangled by overseas markets, they think there is no need to do this. In addition, their leaders like to focus on their expertise and don’t touch other things, so there are good opportunities, but no industrial chain has been developed.
It’s different in mainland China. We might be stuck, so we must develop these technologies ourselves.
Let me emphasize again that the investment in third-generation semiconductors is not very large. The focus is on talent. I personally think that IDM is good, and the division of labor and cooperation with foundries is also feasible.
I hope that people in the investment community will pay more attention and give appropriate support. I would like to emphasize again that the investment money is much less than that of making an advanced logic platform.
The investment amount depends on the corresponding stage
1 billion, 2 billion starting
Q: Can we have a rough estimate of how much the investment will cost compared to advanced factories? Can we have a decent scale?
Zhang Rujing: I personally think that in the third stage, if we only have the materials and epitaxial wafers to make the device, if we use a 6-inch wafer, the investment depends on how big you want to make it. It will probably be at least more than 2 billion. If you reach a scale of 600 to 700 million, you can make money. This is the third stage.
If the second stage is to produce Epi (epitaxial wafer), the investment will not be large, and the corresponding Epi plant investment will probably be less than 1 billion.
The equipment is not difficult, but you need to pay attention to the technology. The raw materials are good from Tianyue and Xincheng in Shandong, China. The materials of these factories depend on how big you want to make them. These furnaces can basically be domestically produced, and they are not bad.
Otherwise, you can buy it from Japan or Germany, and it's not expensive. I estimate that a corresponding factory, not counting the land and plant, will cost about 1 billion to 2 billion RMB to start, and the better it does, the more it will cost. So it's not very big.
China is very strong in packaging and testing.
There is a big gap between the equipment and others
Q: As the founder of SMIC, you have made great achievements in the field of semiconductor manufacturing and witnessed the development of local semiconductor manufacturing over the past few decades. We have made some progress compared to 20 years ago. How do you view the progress over the past period of time? We still have a certain gap with overseas companies in this field. How do you view the gap? In this field, especially in the field of third-generation semiconductors, how can we break through?
Zhang Rujing: I think the gap is too wide. In some areas, China is very strong, such as packaging and testing. As for equipment, such as lithography machines, we have a big gap.
If we specifically look at the materials, manufacturing, design, etc. of the third-generation semiconductors, I personally think that our gap in materials is not that big.
More than two years ago, I visited several domestic companies that made silicon carbon single crystals. The data I saw at that time was quite different from those overseas.
But what surprised me is that in the past two years or so, they have made great progress. In terms of materials, the 4-inch ones are basically very close to those overseas, and the gap is not that big. There is still a little gap for the 6-inch ones, but they will catch up in time. This is about materials.
Epitaxial wafers are completely a matter of technology, and the equipment can also be purchased, so the gap can be narrowed quickly.
As for the third part, the design is not very difficult. It is based on accumulated experience. The manufacturing equipment does not need to be so advanced, but you have to be very careful about the power, otherwise the effect will be different from others.
The difference in effect is quite large, so the focus is still on good design coordination in terms of power.
Because there is currently no real large company in China to do this, I don’t know which one has a strong team developing this.
Therefore, the power silicon carbide products currently produced may be easier to catch up if only the power is considered, but gallium nitride is used for radio frequency.
When it comes to RF, the strongest companies overseas are, for example, Japan's TDK and Murata. We have some gaps with them, but there are also some new companies with talents returning from overseas who are developing gallium nitride in 5G RF, which is quite good.
Please pay attention. I forgot who mentioned it just now, but there are many small and new companies that are doing well. For example, there is a small company called Navitas in California, USA. I don’t know if it has been bought out, but I saw that they were doing well a while ago. There is also a company in France that was bought out by ST and Micron.
As I just mentioned, there are still many good companies in Israel that are worth considering, because what they need most is talent, and they don’t need a lot of talent. If a few good people come and teach our young people, we can almost keep up with them.
We need to consider the talent base in a short period of time. This is one of our weaknesses. We may have laid the foundation, but there is a gap between the foundation and the application. How can we shorten it? European and American companies have done better, so we can learn from their strengths.
So I feel that the gap is not that big. The logic gap is not that big, and the memory gap is not that big. It can be caught up.
I was determined to find the right team. It is very lonely and hard to work in this industry. We need a strong belief to support us, and then we can make it happen. So I am optimistic and believe that we can catch up.
| 
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
