The Millennium Semiconductor Survival Guide.

Chapter 190 Industrial Layout
Compared with Hu Zhengming, whose research fields cover semiconductor components, EDA circuit simulation software, high-end chemical materials in the upstream and downstream of the semiconductor supply chain, and analog IC circuit design, Guan Jianying's areas of expertise are more concentrated in chip design and R&D teams manage.

In other words, Guan Jianying is better at being the manager of Xinxin Semiconductor. Under his leadership, Xinxin Semiconductor can use the dividends of the mobile Internet to become a giant in the field of chip design.

However, Guan Jianying is unable to drive the development of semiconductor companies in the entire park on his own.

And Hu Zhengming can do it. With nearly 30 years of work experience in academia and a member of the board of directors of many upstream and downstream companies in the semiconductor industry chain, he has sufficient ability to judge the overall situation of the industry and enterprises.

For Hu Zhengming, just by looking at people, he can make a preliminary judgment on whether the company has any potential. This is equivalent to data analysis at the subconscious level of the human brain.

Huaguo's traditional physiognomy is also a kind of fuzzy data analysis without systematic induction. In the process of interacting with other people, there will be such a process in the subconscious mind of each person: if in your memory, there is a similar appearance If someone has brought you bad memories, then you will have a subconscious resentment and disgust towards him.

Hu Zhengming is already very skilled in judging semiconductor companies, let alone go to Huaguo to do this. Huaguo's semiconductor companies are still in their infancy, and there is a generational gap between technology and the industry's cutting-edge technology.

From this point of view, Hu Zhengming is the most suitable person to help semiconductor supply chain enterprises in the entire park with technical guidance.

Hu Zhengming smiled bitterly and reminded: "Newman, have you ever thought that if I go to work at Xinxin Technology, you will not have a mentor.

Although you can graduate from Berkeley smoothly if you change your supervisor, you may not be as easy to talk to as me if you change your supervisor.

Newman, you don't want to change your supervisor suddenly after two years of PhD, do you? "

With Zhou Xin's reputation, it is Berkeley who should be worried that Zhou Xin will not be able to graduate normally. Although Berkeley is a public research university and its finances mainly depend on the allocation of the California government, no university thinks that it does not need donations from alumni.

Zhou Xin donated [-] million RMB to Yanda University in the name of student scholarships when he was in China before New Year’s Day. After returning to Berkeley, Hu Zhengming hinted that he should also donate some to Berkeley.

Therefore, Zhou Xin also donated 1000 million US dollars to Berkeley. At the donation ceremony, he also invested [-] million US dollars in the name of Xinxin Technology and the Department of Electrical Engineering at Berkeley to build a laboratory. The final laboratory results were provided by Xinxin Technology and Berkeley. Branch shared.

In any era, 2000 million US dollars is not a small number. In 4.95, the NNI program allocated to the nanomaterials research program at the national level by American only allocated [-] million US dollars for the whole year.

The $4.95 million is still split between Oak Ridge National Laboratory, Berkeley National Laboratory and Sandia National Laboratory.

Of course, it’s not bad for Zhou Xin, who can use Berkeley’s professors and research dogs to help him do some projects that he wants to do but doesn’t have the energy to push forward. Anyway, the final scientific research results are half of Xinxin Technology.

As for whether American will recognize it or not in the future, we will talk about it when they do not recognize it. At least American will not be able to recognize it in China.

After these two incidents came to light, Berkeley became more enthusiastic about Zhou Xin, wishing that Zhou Xin would stay in Berkeley for a few more years and then graduate smoothly.

Before Zhou Xin, the famous alumni of Berkeley were concentrated in the field of scientific research, and there were not many alumni in the field of Internet entrepreneurship, which could not be compared with Stanford and MIT.

With Zhou Xin, UC Berkeley has a significant alumni in the Internet field.Of course, this can only serve as icing on the cake. The most important thing is the resources at Zhou Xin's hands, which determines that Berkeley will guarantee Zhou Xin's smooth graduation in Berkeley no matter what.

Zhou Xin said: "Professor, even if you go to work in Zhangjiang, you can still be my supervisor at the doctoral stage.

Berkeley can reserve a teaching position for you, and you can guide my academic work remotely while you work in Zhangjiang. "

According to the original historical trajectory, Hu Zhengming went to TSMC to work as the chief technology officer for three years, and the professorship of the Department of Electronic Engineering at Berkeley has been reserved for him.

It's 2001, not 2021, and jobs in academia aren't that tight yet.Moreover, with Hu Zhengming's qualifications and academic achievements, even in 2021, Berkeley will reserve an academic position for him.

Hu Zhengming said: "I still need time to think, mainly because I am too old, going to Zhangjiang is equivalent to opening up wasteland from scratch, which is a very big challenge for me.

You know, I don’t like to take risks, otherwise I wouldn’t have worked in Berkeley for the rest of my life. Even if I was doing Silue Semiconductor, it was a small-cost startup.

To be honest, compared with entrepreneurship in the fields of circuit design, chip manufacturing, semiconductor materials, and semiconductor equipment, the start-up cost of EDA software is the lowest.

Circuit design requires engineers of different modules, and then it needs to be tested and taped out. The final designed chip needs to be paid for by someone after it is produced, and the cost is very high.

Not to mention chip manufacturing, any production line is expensive, and the same is true for semiconductor materials and semiconductor equipment.

Take the lithography machine as an example. 20 years ago, at least US$1 million was invested in the research and development of the lithography machine. At least five hundred million dollars should have been spent, right?
Otherwise Nikon would not sell it to you at all.

5 million US dollars is enough to think about semiconductor operations for ten years.

So different people have different preferences for risk. I am a relatively risk-averse person.

In the business of Silue Semiconductor, the main business of Silue Semiconductor is to make EDA software, and at the same time, it will do some foundry of chip design.

Chip design foundry also serves EDA software. If we don't use Silue EDA software ourselves, how do we know how to optimize it?

We invested a total of 500 million US dollars in Silue Semiconductor when we first started our business, of which I didn't pay any money, and invested purely in technology.

Of course, I believe that I will go to Zhangjiang with you, and you will not let me pay, but there are too many uncertainties in Zhangjiang.

Judging from my experience of giving lectures at Yanda for so many years, Huaguo is very short of talents in the field of integrated circuits.

I probably started in 1995, and every year I went to Huaguo to do some academic exchanges, and then gave some academic lectures to Huaguo students.

Yanda should gather the top students in Huaguo. They are very talented and smart, but they lack practice. In Berkeley, you can enjoy the resources of Silicon Valley, you can practice in Silicon Valley, and you can learn about the front-line industry in Silicon Valley. What, what kind of technology will help them find a job or start a business in the future.

Students in Huaguo only have this concept after they graduate, and when they were in school, they didn't know what the most advanced technical direction and technical route in the industry were, and then they didn't know what kind of work they liked to do. After spending many years groping, when you figure out what you want to do, you may not have the state of mind to start from scratch.

I have always felt that there are problems in Huaguo's talent training mechanism.

Of course, having said that, for integrated circuit entrepreneurship in China, especially as you mentioned, extending investment in upstream and downstream semiconductor companies around chip foundry, lithography machines, and chip design, the talent gap is even greater.

The difficulty in the middle is not 01:30. To really do this when you go back, you need to be patient. This will be a protracted battle.

At least three to five echelons have to be cultivated in talent cultivation, which will take five years, and then the results can be produced in a five-year planning cycle. I estimate that some results will be slowly produced in the second five years. "Hu Zhengming spoke very sincerely, and what he talked about was his thinking around the current situation of Huaguo.

Unlike Zhang, Hu Zhengming has feelings for the mainland and has a psychological existence of wanting to support the development of the mainland, but because of his risk preference, he can't give up everything he has now and return to China for development.

Hu Zhengming went to Wanwan when he was a baby, while Zhang spent his teenage years in the mainland, and did not leave the mainland to study at Harvard until he was 49 in 18.

As for the position, Zhang Zhongmou's attitude on the surface is to always maintain neutrality and not support any side.

However, judging from the results of TSMC’s construction of the factory, it finally stood on the side of American. Originally, TSMC only built factories with advanced processes in Wanwan, and foundries with mature processes were located outside of Wanwan. The advanced technology factory was moved to American.

Zhang Zhongmou's attitude can also be inferred from the results in reverse, because although he resigned as the chairman of TSMC, he split the power into two when he resigned and handed it over to Liu Deyin and Wei Zhejia respectively. Holds decision-making power on important agendas.

When TSMC built a factory in American, Zhang Zhongmou publicly stated that even if the manufacturing cost of building a factory in the United States is 50% higher, TSMC will continue to transfer some production lines to the United States.

Zhou Xin didn't know that Zhang Zhongmou would stand in public in 2022, but this did not affect his understanding of what kind of person Zhang Zhongmou was from Hu Zhengming when he was a doctor under Hu Zhengming.

Of course, I know more about Zhang Zhongmou from Liang Mengsong. Liang Mengsong has worked in TSMC for many years after graduating from Berkeley.

Zhou Xin said: "Professor, I can fully understand your concerns. I also have this patience. My funds are sufficient. Let alone ten years, 15 years is fine.

Moreover, Huaguo has a vast market, and the local market is enough to cultivate some excellent semiconductor companies. It is not absolutely a loss to do this.

Part of it can be profitable, such as chip foundry, the processor of the Matrix mobile phone is designed by Xinxin Semiconductor, and we also designed a decoding chip for Apple. These are all income.

With these two items alone, it is not a problem to support Xinxin Semiconductor.

The same is true for other areas of integrated circuits. Being backward means that the company is small, and a small company means that the profit-loss requirements for achieving breakeven will not be too high.

So what Zhangjiang needs to do is to build up the ecology. After the ecology is built up, a positive cycle is formed. I believe that with the wisdom of the Hua people, this matter can be done well. "

Later, why Huaguo Semiconductor developed slowly, a large part of the reason lies in the upstream and downstream of the chip industry. Whether it is a branch of a foreign company in Huaguo or a local company in Huaguo, they are unwilling to use the products of Huaguo local companies.

They are unwilling to give local companies in Huaguo any chance.

This is because there was no chain owner enterprise with a core position to play a leading role at the beginning.

Zhou Xin continued: "I have also communicated with Huaguo regarding the issue of insufficient talents.

Starting this year, Zhangjiang Hi-Tech Park will sign cooperation agreements with several major universities in Shenhai to provide students with internship opportunities. The internship will last from their freshman year to their senior year. We have also discussed internally that for students who perform well, Came here to work directly after graduation.

This plan is not only for Shenhai students, but will be open to the entire Huaguo 985 colleges and universities in the future.

The current problem is that there are not enough companies, and the number of internships that can be given is only enough to satisfy Shenhai University.

Professor, you have been in college for so many years. I invite you to Zhangjiang because you have a lot of experience in the field of industry, education and research and can play a big role in this area.

The internship plan we have developed is very superficial. How to achieve the effect in the specific implementation process, I think you can plan it, and it will definitely have better results. "

Zhou Xin wanted to poach Hu Zhengming to Zhangjiang, and Liang Mengsong was also stable in the follow-up.

When SMIC became chaotic, he entered SMIC forcefully, then took Liang Mengsong in to take over SMIC, and kicked out all those who wanted to make trouble.

Due to the scattered ownership of SMIC in later generations, internal fighting was quite fierce, which gave many people the impression that SMIC was good at internal fighting.

Of course, SMIC is really good at internal fighting, and the level of internal fighting is comparable to Gongdouju.

Hu Zhengming said: "You still need to give me time, I need time to think."

"Uncle Lin, it is because of your Xinxin lithography machine that we can get on the right track so quickly.

The first prototype came out much faster than I expected. "Zhou Xin personally ran to San Francisco International Airport to pick up Lin Benjian.

This time when Lin Benjian returned to American, one was to visit relatives. His wife went to Zhangjiang with him, but neither of his two daughters followed.

After shaking hands with Zhou Xin, Lin Benjian said: "The engineers in Huaguo are very diligent, and their diligence is no less than that of me who just went to IBM to work.

Although they lack experience, they work very hard and learn very quickly.

With their hard work, Xinxin Lithography Machine can produce the prototype so quickly. If it is not for their cooperation, I will not be able to do this alone no matter how good I am.

Newman, I haven't seen you for a long time, it should be a whole year.

When I came back from Christmas vacation last year, I only stayed at home for two weeks, and we didn't have a meal together at that time. "

Zhou Xin nodded: "Yes, I came to see you this time because we haven't seen each other for so long, and there is another important reason.

Before I talk about this reason, I would like to give you a gift to congratulate you on your initial achievements in Zhangjiang. "

After Zhou Xin finished speaking, he handed a box with the Matrix Logo to Lin Benjian: "Mphone, if you have followed Silicon Valley news recently, you should know about this product I designed."

Lin Benjian took it and said: "Of course, I often see news about Mphone on Huaguo's online forums and portals.

As time goes by, more and more people begin to admit that Mphone is an epoch-making product.

But I heard that since the servers of Mphone's built-in applications are all set up in American, the experience of using Mphone in China is particularly bad, and its essence can only be experienced by one-third in China.

Because the network delay is too high.

Doesn't have the silky smooth feel of the American. "

Zhou Xin said: "Yes, we have been talking with the Ministry of Information Industry of Huaguo recently about the entry of Mphone into Huaguo.

I guess it should be soon, and it will enter Huaguo by the end of this year. The server environment of these applications has already been deployed in Huaguo. "

Zhou Xin turned the topic back to the surprise mentioned just now: "I have almost talked with Cymer here, and they are willing to accept the acquisition of Xinxin Technology."

A smile finally appeared on Lin Benjian's calm old face: "What do I need to do when I come back this time?"

Zhou Xin said: "Dr. Richard wants to talk to you, he still has many doubts about the future.

I don’t know much about light source technology. You already know Richard. We need to tell Richard why Xinxin Semiconductor acquired Cymer, and what we will and will not do after the acquisition of Cymer.

Dr. Robert and I have talked very clearly before, Dr. Richard may want to hear you again. "

Cymer, a company that makes light sources. The light source here is the light source used in lithography machines. Cymer is in a leading position in technology, and its market share exceeds 70.00%.

The light source is one of the most critical technologies upstream of the lithography machine.

The light source and the lens together form the source of light for the lithography machine. Compared with Zeiss, the giant in the lens field, Cymer is still affordable.

Zeiss really can't afford it.

With a market value of only 30 billion, Cymer is completely affordable.

In fact, Cymer is very important and has always been a leader in technology. The market in the field of light sources is very small. It occupied 2000% of the market in 80, and its revenue was only 5 million US dollars, which belongs to the leader in the segmented field.

Originally ASML would acquire Cymer in 2009, but in 2000, when Lin Benjian and Zhou Xin were talking on the phone, Lin Benjian saw the news of Zhou Xin’s acquisition of Nvidia and ARM in the news, and proposed that if there was an opportunity, he could consider acquiring Cymer. .

In Lin Benjian's view, Cymer's technical strength is very strong, and the acquisition of Cymer is conducive to their technological development, which is a strategic acquisition.

After learning about it, Zhou Xin verified Cymer's information and memory, so he directly made up his mind to acquire the company.

It turned out that it was very difficult to talk about it, because Richard and Robert, the founders of Cymer, both had PhDs in the laser field, and later started a business together to focus on making light sources.

Their entrepreneurial process has been very smooth, and they were successfully listed on NASDAQ. The bursting of the bubble did not have a great impact on this monopoly.

It is nothing more than a drop from 50 billion to 30 billion.

In the second half of last year, Zhou Xin chatted with Robert on a weekly basis, about future development, and about their desire to reshape the pattern of existing lithography machines.

Robert was fooled, and there was another founder, Richard. Compared with Robert, Zhou Xin's halo had no effect on Richard.

Richard just wanted to chat with Lin Benjian.

As a light source researcher at IBM, Lin Benjian knew Richard and Robert, and was even more familiar with Richard. They often chatted for a day at the International Optical Conference.

Lin Benjian said: "Richard doesn't like change very much, he would be more repulsive about being acquired.

Even if you use a higher premium and promise that the other party will not interfere with Cymer's business, Richard is not willing to accept it.

Since Richard is willing to talk to me, it means that he has been persuaded by Robert, and he just needs me to analyze it for him as an objective third party.

You can leave it to me with confidence.

For us, the acquisition of Cymer and the integration of Cymer into Xinxin Technology's system are of great significance. I think I have found a solution for overtaking on curves in the field of lithography machines, but this solution is not enough for me. hidden.

The lack of concealment here means that this plan is easy to leak. After the leak, others can follow our plan and catch up in the manufacturing process.

If there is Cymer, we can use Cymer to interfere with their judgment.

Mistakenly tricked Nikon, Canon and ASML into thinking that we had adopted Cymer's most advanced technology and let them focus on the light source. "

After Zhou Xin listened to it, he guessed that Lin Benjian must have thought of the immersion method, which is to use water as a medium between the lens and the light source to reduce the wavelength.

The refraction ratio of water is 1.44. The original 193nm wavelength of light is reduced to 134nm after refraction, and the corresponding manufacturing process is also greatly reduced.

Relying on the immersion method, the chip manufacturing process has evolved from 45nm to 7nm in just ten years.

But it is true that once the immersion method is leaked, others can quickly catch up.

In addition, American will naturally support ASML, which is backed by Wall Street capital, instead of developing and producing new core lithography machines in Huaguo.

If there is no technical advantage, or if the advantage is not obvious enough, the chip foundry will not switch from the previous lithography machine to the new core lithography machine.

Even purchases are made in small quantities.

Lin Benjian continued: "However, using Cymer as a cover-up method won't work for long, at most one or two months, Nikon and the others will be able to discover this secret sooner or later.

But this is not important. The acquisition of Cymer will allow us to use Cymer's most advanced technology in the subsequent research and development process, which will help us catch up with Nikon in the field of lithography machines. "

Zhou Xin said: "Everyone competes fairly. I am not so anxious to make the lithography machine reach a certain level. As long as it can be used, it is good to slowly catch up with the advanced manufacturing process.

However, there are companies like Cymer, you can tell me that if it is within my tolerance, tell me as soon as possible that the capital market has not yet recovered, and it is cheap to buy.

If it was the end of 1999, buying Cymer at that time must have made my heart hurt. "

(End of this chapter)