Welcome to the “Genesis” Wechat Subscription Number: sinachuangshiji
Source: deardata (ID: deardata)
I met Professor Wang Zhihua again at the Institute of Microelectronics, Tsinghua University. Every time we enter the East Main Building of Tsinghua University, there is always a feeling of space-time reversal. Compared with the brand-new office building full of decoration materials, the glossy terrazzo floor makes us step into the first few episodes of the TV drama “River”, back to twenty or thirty years ago. In Professor Wang’s office, a wall was completely filled with paper materials. The color and style of office furniture are the latest in the 1990s. The only trendy item is the white stand-up platform on the desktop, which allows Professor Wang to stand on his computer.
The expression on the professor’s face is still very serious and rigorous. A few simple greetings, and soon the discussion got to the point. China’s integrated circuit industry is undoubtedly facing a new round of situation, and the issue of China’s chip talents is worth discussing. To put it more highly, that is, when AI and semiconductor have become the high priority development strategy of the country, chip talents should support the strategy, the issue of talent is worth discussing, and we should listen to voices from all walks of life, such as obstetrics.
The background of integrated circuits is always strong. There always seems to be a voice that chips in our country are scarce. So what does Professor Wang think of this problem?
From 7% to 50% of global output, then we will be short of 350,000 to 800,000 people.
Wang Zhihua is deputy director of the Institute of Microelectronics, Tsinghua University and professor of microelectronics and nanoelectronics, Tsinghua University. As a scholar born in the 1950s, Professor Wang Zhihua has been in Tsinghua Garden for more than 30 years. Professor Wang is good at scientific research, occasionally explains the industry and popularizes integrated circuit knowledge. Professor Wang also sends microblogs to express his views. He joked that his microblog had 50,000 chip fans. Asked about the origin of Weibo’s name (@Wang Zhihua Mo Ping), Professor Wang looked at me, smiled in two lines and said, “I’m from Simping County, Shandong Province.”
In Professor Wang’s opinion, when it comes to industrial talents, we must first talk about industrial scale. I named his calculation method “Inverse Reasoning Theory”. Estimate the manpower and workload needed by the industry according to the target. It’s how much work to do, and how many people to do it.
I asked, “What is the goal?”
Professor Wang said: “The market is the goal.” He further explained that the current scale of the global integrated circuit industry is the goal of the “realistic version”.
He produced his own “global integrated circuit industry scale structure map”, which is often used as a necessary “courseware” for public speeches. He stressed to me, “If you want to write chips and read global integrated circuits, you need to understand this picture first.” Professor Wang’s eyes were bright, as if the teacher on the platform was afraid that the students could not grasp the knowledge points.
“Look, this is an inverted triangle. At the top of the tower is China Science and Technology Corporation, the world’s highest market value and the first echelon, which is the most proud achievement of Chinese science and technology practitioners. Here are the names of companies like Ray Guaner: Alibaba, Tencent, Baidu, Mei Tuan… The monetary unit of this output value is Trillion (trillion), and the telecommunications industry next to it is also trillion units. Further down, as the cornerstone of high-tech industries, the unit of currency calculation is Billion.
“The brilliance of the upper layer cannot replace the foundation layer. Faced with different battlefields and experienced different periods of development, if we start to talk about the basis of industrial development, that is another story. Professor Wang stopped talking about the history of industrial development and stopped talking about it. That is another topic indeed. Although it is also the historical background of the whole chip industry, the task of this topic is to talk about chips.
“Since it’s an inverted pyramid structure, the stronger the foundation, the more stable the structure,” he said.
This chart tells us that in 2017, the global scale of the information industry chain supported by integrated circuits and software was about 420 billion US dollars, and that China’s integrated circuit products were about 30 billion US dollars (7% of the world). But that year we consumed, that is, imported 60% of the world’s integrated circuit products.
“Fewer production, more use, this is the status quo.”
Professor Wang did not comment too much on this seemingly unsatisfactory report card. His face was heavy and his office was quiet after a pause. Obviously, he closed the conversation.
He said: “The reason is simple: how many lives, how many people. Set targets according to the current situation, and then check the target. Professor Wang quickly shifted his focus to “how to do it”.
The first step is to set a target and estimate the total demand for talents from the target of gross output. If our country wants to develop into a large integrated circuit country, it must occupy at least one-half of the global share. One-half of the $420 billion is $210 billion. Calculate how many people can make so many IC products with so much output value.
The second step is to look at the per capita output rate of integrated circuits in the United States. We need to look at the top five companies in the industry. They are Broadcom, Intel, Qualcomm, Meguiar Semiconductor and Adeno Semiconductor (ADI). In this picture, the vertical axis is in millions of dollars and the horizontal axis is in years, describing the development of the past decade. Broadcom, the highest per capita output producer, has an annual output of $620,000. Intel, which has the second highest per capita output rate, has $520,000 per person per year. By contrast, Adeno Semiconductor, an analog and digital hybrid circuit designer with complex circuits and small sales, produces less output per capita, at $310,000 per person per year. Compared with the US transcripts, US companies with high productivity need 150 people (refer to Broadcom) and those with low productivity need 300 people (refer to Adeno Semiconductor).
Professor Wang Zhihua’s calculation results come out. “If we aim at $210 billion, we need a team of engineers and technicians of 350,000 to 800,000 people.”
Professor Wang worked out the answer to the number gap. There is a large amount of information in the calculation process. I took notes, needed to go back and digest them, and asked Professor Wang for his PT.
At the same time, Professor Wang Zhihua recommended to me the White Paper on Talents in China’s Integrated Circuit Industry (2017-2018). Data show that by the end of 2017, the number of employees in the integrated circuit industry in China is about 400,000. Around 2020, the talent gap will reach 320,000 people.
“We are working towards the goal of a large integrated circuit country, which is to plan for a 10-year cycle. Although it is not necessary for more than 300,000 engineers and technicians to be on duty immediately, the growth rate of personnel training has not yet reached the target. At the present speed of personnel training, the number and demand of personnel training in the next ten years still do not match. Take Peking University and Tsinghua University for example, there is less than one class per year for undergraduates in North China University of Qinghai. In recent years, 10 to 20 undergraduate students have been trained each year to meet the quality standards of personnel training, and 100 to 200 undergraduate students have been trained in ten years. Other IC brothers have trained hundreds of students in scale and thousands in ten years. At present, dozens of colleges and universities in China can not meet the demand for talents in the integrated circuit industry, and the total amount of personnel training is not optimistic. The supply of industrial talents does not match the growth rate of industrial development. Professor Wang Zhihua’s face became more and more serious.
Taking Peking University as an example, I checked the specific data. The bad news is that according to the statistics of Peking University graduates, there are 33 graduates of microelectronics science and Engineering in 2018. There are 37 Masters of Integrated Circuits with a small number. But the good news is that fewer students correspond to abundant teachers and more professors are allocated to each student. The students who have been admitted to Peking University have indeed obtained very good educational resources.
The current situation is that relying solely on Colleges and universities to train talents can not meet the needs of industry. According to statistics, the proportion of graduates entering the industry in the field of integrated circuit specialty is relatively low, and only about one tenth (12%) of graduates enter the industry for employment. About one-third (36%) of the graduates from the four major fields strongly related to integrated circuits enter the industry. The major of integrated circuit is: graduates of microelectronics science and engineering, microelectronics and solid electronics, integrated circuit design and integrated system, integrated circuit engineering.
When it comes to the second point, the smile on Professor Wang Zhihua’s face completely disappeared. “However, in this unoptimistic context, we have also lost some IC engineers and technicians to the financial and software industries. It’s not just about pay. Because salary is not the only determinant of talent retention. Nowadays, many companies such as Huawei Haisi can offer market prices with annual revenue of 300,000 yuan. Professor Wang Zhihua thinks this is a reasonable price. The phenomenon of Pay differences among industries has always existed. Enterprises and universities should continue to observe and study the core causes of brain drain to retain talents for the industry.
Salary is a sensitive survey. For this reason, I specially found a graduate Master of integrated circuits in microelectronics major of Peking University in 2019, and inquired about the students’graduation prospects.
According to the student’s introduction, Huawei Haisi accounts for about 40% of the students’whereabouts. Some go to Baron to make 5G chips and others to Taishan. Some go to Fudan Microelectronics Company and other state-owned enterprises, others go to Texas Instruments, Intel and other foreign-funded enterprises. Some students choose BitContinent, Cambrian and other start-up companies. Some students continue to study for doctoral degree. Of course, some students have lost to enter the financial (banking) and software industries, and did not choose the integrated circuit industry. He specifically disclosed that Huawei Haisi Semiconductor Company can offer 280,000 to 330,000 annual salaries for the current master’s degree in Peking University, depending on the differences of students’negotiation ability and personal competence.
At present, relying solely on Colleges and universities to train talents can not meet the needs of industry. According to the survey of White Paper on Talents in China’s Integrated Circuit Industry (2017-2018), the average monthly salary of the industry is 9120 RMB, ranking sixth in 52 industries. From a global perspective, China’s integrated circuit talent cost advantage is obvious, there are value depressions for excavation. However, the IC industry has a long cycle and the talent growth cycle is relatively long. In addition, there is still a significant gap between salary treatment and Internet, finance and other industries. Because of the high time cost and income gap and other factors, the attraction of talents in this industry is seriously affected.
When Professor Wang Zhihua talked about this, the light in the office seemed to be dimmed. Professor Wang needs to put more effort into explaining the problems of talent and market development. He took a sip of the green tea in his cup.
Professor Wang Zhihua concluded that in terms of market issues, it is crucial to recognize the two markets, stock and increment, and pay attention to them separately. Because we need to meet the demand of talents in both markets simultaneously. It needs not only a group of well-skilled IC craftsmen in the stock market to conquer the products originally conquered overseas, but also innovative talents to solve the unsolved problems in the industry and to take innovation as a cutting edge to enter the incremental market. Both markets need to be fought. The integrated circuit industry has high demand for talents and great technical difficulties. Under the objective situation of loss, people should be eliminated. Ultimately, we should ensure that there are enough engineers and technicians who can develop healthily.
Looking at my dignified face, Professor Wang Zhihua took on the role of adjusting the atmosphere of dialogue. His lips were raised and his eyes were crooked with laughter. He said, “Talents in an industry should not only come from the training of colleges and universities.” We should be optimistic, as long as we are willing to invest, if there is not enough talent, we will go to the relevant industries to dig, but also to the United States Silicon Valley to grab.
Chips want to succeed in business, which is not only a technical problem, but also a commercial problem. Market, talent and capital are indispensable. Throughout the history of the global semiconductor industry, there are many stories about Silicon Valley in the United States. Professor Wang reminded me to read several chapters of Wu Jun’s Tide Peak. Meditation, with the participation of capital, this period of history has the meaning of “habitually looking at the autumn and spring breeze, turning the tables for success or failure”. From Wu Jun’s point of view, Wall Street can not only make semiconductor companies shade the sun in the industry, but also make it sunset in the west. To some extent, semiconductors are aristocratic games.
Past is gone, look at the present.
Over the past two years, on the one hand, the state and local governments have given strong support to funds and policies, on the other hand, the enthusiasm for capital has not diminished. Semiconductor momentum seems to be sunny. However, there is a light, there is a backlight naturally. At the hottest time when the Internet is burning money, the business model has not yet come out, the cost of manpower has risen first, and practitioners directly benefit from “as long as they change jobs, they will get higher salaries”. Whether this phenomenon exists in the chip industry and how to treat it, I found a partner of Arctic Lights Venture Capital, Mr. Yang Lei, at the Summer Summit of Anchuang Future Science and Technology Creation “Core” Conference in 2019.
Arctic Lights Venture Capital Partner, Yang Lei: Driven by innovation, China’s chip industry lacks high-end talent and excellent team.
Unexpectedly, Mr. Yang Lei and Professor Wang Zhihua’s “reverse reasoning theory” have the same thinking, they all look at the goal first. In his original words, “We will wait and see how many listed companies can be created by targeting hundreds of billions of markets.”
“The enthusiasm of China’s semiconductor industry contrasts sharply with that of the United States on the other side of the ocean.” Arctic Lights Venture Capital Partner Yang Lei told me that semiconductor investment in the United States can almost be counted by hand, maybe not one hand, so many investors back and forth. From the age of the team, in fact, the age of semiconductor workers in the United States is generally aging, while China’s young power is worth looking forward to, there are a large number of new technological talent force is accelerating growth.
On the issue I raised, Yang Lei expressed his concern: “The flow of high-end talent is not necessarily driven by complete rationality, but also needs a lot of financial support, commonly known as money-burning digging. Such a situation is bound to lead to a sharp rise in the cost of talent. For practitioners, wages are rising at present. However, because industry resources are not concentrated in important areas, such growth can not last long. There are too many stories about the last Internet growth cycle.
“The cost of talent in China is rising rapidly, and most investors in China are not mature enough, so bullets will be scattered. At this time, our limited resources will be scattered, which will lead to the rise of talent in the market competition and reduce our competitiveness. This is indeed a concern.” He said frankly that was a problem. Yang Lei emphasized that there is no chance to follow the international integrated circuit products without innovation. Unlike the Internet industry, the technology industry has relatively obvious regional characteristics. The same products are better abroad. Why do you have to buy your local ones? So innovation is inevitable. He also wanted to express to his colleagues in IC investment institutions, “Don’t invest in low-end. What China really lacks is an excellent team. In fact, if you can’t meet a particularly good team, you don’t have to invest in it.”
Chief Executive Officer of Guangjian Technology, Tsinghua University, Doctor, Berkeley, California, USA: Zhu Li
If this wave of semiconductor development in China has attracted many new chips with dreams to leave Silicon Valley, Zhu Li, who graduated from the Department of Electronic Engineering of Tsinghua University and PhD from Berkeley, California, is one of them. Former director of deep camera module design at Apple, he returned to China from the United States in April 2018 and is now the CEO of Guangjian Technology. One year after its establishment, Guangjian Science and Technology has formed a team of nearly 70 people, more than 80% of whom are R&D personnel. He said that people of our age range from 85 to 90 are now returning home to start businesses in particular. Most of them have returned abroad one after another. Chips are the choice of the country, which means that resources such as policy investment, investors, large companies and so on will go in this direction, and participants will have the opportunity to grow rapidly.
It is a reality that there are gaps in talents, but in the long run, where do they come from? From the United States, the cultivation of integrated circuit majors in Colleges and universities is the path. So, with engineering background, are there opportunities for non-integrated circuit related professionals to take part in the integrated circuit industry? With this in mind, I met an engineer from Corehara Microelectronics (Shanghai) at the Shanghai Station of CES Asia Expo in 2019.
Corehara Microelectronics, Chip Solution Engineer: Li Bei
As a post-80s semiconductor man, he is full of a positive and optimistic vitality, we can see that he is very fond of his post and work. After introducing the business of the company’s chip design, platform service provider, he was happy to share his professional experience.
“The answer should be feasible,” he said without hesitation.
Because he himself is an example of switching from an engineering major to an integrated circuit. After graduating from Hunan University, he worked in ZTE Communications for ten years, starting from the field programmable gate array (FPGA) research and development engineer, and has been a chip project engineer.
Seeing that I care so much about the talents in semiconductor industry, he also spoke frankly. “This industry has a general salary and a small number of people. Chip companies are not as well-known as large enterprises in the Internet industry, and their posts are not as bright as they are. But semiconductor people have fun.” He also introduced the direction of the industry: “If other engineering specialties want to enter the semiconductor industry, we can start with the design of digital chips, each chip factory has a complete design process, and each link is very subdivided, with a certain digital logic basis and hardware programming language basis can be.”
There are still many problems about talents that need to be discussed in detail, at the end of the article. I’ll add two more.
“How long will it take to train a group of chip talents?”
Yang Lei, partner of Arctic Lights Venture Capital, said that good semiconductor talents need to go through product cycles, which may take five or ten years.
I asked, “How can we calculate it in five years?”
He answered that because the basic development cycle is two years, we have to sell it to the market for two years, and then spend time observing and feedback. So, basically five years can be a small success, and ten years can be a big success. He emphasized that it was difficult for a semiconductor talent to make something without 15 years in the industry. Integrated circuit is a long cycle, cold bench, talent should be accumulated slowly. This long process is fair to any company. Once you build it, no one else will. This is the moat.
“What is the talent structure of integrated circuits in China?”
Professor Wang said that the talent structure of integrated circuits is very special, and the talent structure of analog circuits is basically flat. The reason is that there are many kinds of products, hundreds of them. You know what? The number of products on the product chain of ADI and Maxim is larger than that of engineers. The digital circuit is different. Talents are divided according to the type of work. Some write logic code, some structure code, some do layout and wiring, some do simulation, and some do verification. Every level of work cannot be substituted for each other. The type of integrated circuit work is divided from the product chain, and the type of work itself is not divided from the level of technology. Only different, no high or low. I thought to myself, generally speaking, “365 lines, travel first prize”, the characteristics of integrated circuits is “every type of work out first prize”.
I asked, “Does the high content of integrated circuit technology mean that it is difficult to train talents?”
Professor Wang shook his head and waved his hand slowly. “Enough incentives and salaries, increase investment in scientific research, and bring outstanding talents to the IC industry.” I don’t know why, I always feel that his last two sentences have a great echo left in the corridor of Tsinghua East Main Building, and also in my mind.
His views are clear and his attitude is optimistic. Although today’s topic is not easy, there is no mention of breakthroughs and progress similar to “fierce” and it seems that they are all looking for problems. However, at the end of the conversation, I can feel Professor Wang conveying to me in the atmosphere of “simplicity of Shandong” and “practical work of Tsinghua”. He should not only be open-minded in reality, but also full of confidence in the future.
At the end of the conversation, walk out of the main building. In my arms is Professor Wang’s new book, Brief History of Circuits and Systems. In the dusk, there was a light rain, the thunder was faint, the pages were fragrant and the wind was blowing. I didn’t want to go. I stopped in Tsinghua Garden, and all the scenery I saw was shrouded in the gentleness of midsummer. Whether it’s ear or brain, there are a few low baritones in my mind: “Don’t listen to the sound of trees beating leaves, why not whistle and walk slowly?” Who’s afraid of bamboo sticks and awn shoes? Looking back on the bleak places, returning, there is no rain or sunshine.