An article to understand the localization of semiconductors
Latest update time：2020-02-23
Reads：
Investment warning: The author has no interest in the target companies or investment decisions involved in the report. This series of articles is only for personal interest. Investors should not use this report as the only reference factor for investment decisions. If there is any investment loss due to this article, the consequences are at your own risk. At the same time, experts who are interested in the field of science and technology (such as semiconductors, chips, AI, data, etc.) are welcome to privately mark the backstage and communicate with each other.
introduction
Let me tell you why I am in this neglected place, the science and technology section.
First, in interviews with PE, VC, industry research, etc.,
a question they like to ask is, what are you paying attention to recently/please analyze the sectors and investment points that you are familiar with.
I chose an industry with high barriers to entry - the electronics sector - for analysis.
Second, there are many technical and hacker masters around me.
I would rather take the opportunity to hug a technical dad than miss it. (And if you study it in depth, you will find that some research reports on the market also make mistakes)
Third, hobbies.
I have read n research reports, n roadshow PPTs, and various financial data on Wind... I am full of interest in chips, semiconductor materials, and semiconductor equipment, and I look forward to the arrival of "localization of science and technology".
The era of intelligent connection based on 5G is coming, so it is good to follow the trend. Otherwise, how can I be the most fashionable and handsome dog when I don’t understand cloud computing, big data and blockchain.
The investment points in each era are different. The key is to cultivate your own sense of smell and ability to see problems in this ever-changing landscape. For example, PE equity investment in 2012, P2P in 2014, mobile Internet apps in 2015, the Internet celebrity economy in the past two years, and the future big health care and artificial intelligence.
However, investment should never be made in areas that you are not good at, so we should not blindly grasp the so-called hot spots, like the guys who were keen on investing in Bitcoin a few years ago, who did not even know the principles and risks behind Bitcoin, and kept talking about how great Bitcoin is and how high-tech it is, but now they have become victims of capital. So I hope that friends who are paying attention are really interested in this, or really want to cultivate their ability to look at problems.
Many people on Weibo and WeChat public accounts said that they don’t like reading boring research reports, but they want to enhance their investment sense. So we won’t follow the traditional “boring research report style” here. There are many professional terms in the electronics sector. In order to make everyone feel relaxed and rewarding, this article strives to explain the logic in a relaxed style.
The Dilemma of Chip Localization
A senior who works in PE told me an interesting story about how they entertained cadres. The second-in-command cadre of a provincial capital city actually thought: "If there is a shortage of steel, build a steel plant; if there is a shortage of cement, build a cement plant; if there is a shortage of chips, build a chip plant."
At that time, I didn’t understand the deep meaning of it. I simply thought that since chips were China’s development point, I should just invest and build it.
If chips cannot be localized, build chip factories; if semiconductor materials and equipment cannot be localized, build semiconductor equipment factories. This path is far from enough.
(China has plenty of funds and investment institutions such as PE/VC)
There is still a long way to go for the localization of chips, and it requires more power injection
1. Why do we need “localization”?
Localization is multi-factorial.
On the one hand, for security reasons:
The Prism program implemented by the United States.
Since 2007, the National Security Agency and the Federal Bureau of Investigation launched a secret surveillance project code-named "Prism", which directly entered the central servers of Internet companies to mine data, which aroused our concerns about security.
On the other hand, due to price and supply and demand factors:
They are all imported from abroad, and the prices are relatively high. And Trump suddenly changed his mind and stopped the supply. How can Chinese companies solve this thorny problem?
Therefore, domestic substitution is the general trend. Otherwise, China's development will always be controlled by others.
2.
Difficulties faced by localization
First of all, let’s clarify one question: what needs to be localized?
The semiconductor industry chain can be divided into upstream, midstream and downstream
Upstream: that is, the materials and equipment required to prepare semiconductors
Midstream: Use equipment and raw materials to prepare semiconductors. [Including integrated circuit design, integrated circuit manufacturing, integrated circuit packaging and testing]
Downstream: Use integrated circuits in automobiles, consumer electronics, etc.
Which step is difficult for us to localize?
Both!
The difficulty of each link varies.
Downstream, such as the CPU (central processing unit, explained in Chapter 1, mainly controls the time and space of the instruction flow. The main performance of the computer depends on this. If the host is compared to the body, then the CPU is equivalent to the brain)
If we want to localize CPUs,
is it enough for us to focus on improving chip performance parameters (such as main frequency)?
Main frequency: The clock frequency of the CPU. The operation of the computer is performed step by step under the control of the clock signal, and one step of operation is completed in each clock signal cycle. Therefore, the high or low clock frequency reflects the speed of the CPU. There is a certain relationship between the main frequency and the actual operation speed, but it is not a simple linear relationship. This also means that the operation speed of the CPU also depends on the performance indicators of various aspects such as the CPU bus. Anyway, remember that the main frequency is an indicator to measure the operation speed of the CPU.
No, it is not enough. At present, domestic chips still lack a complete CPU ecosystem like foreign ones. (CPU chips are important, but the supporting operating system and supporting applications are also important)
For example, domestic computers can't even use the software Photoshop, so what should the beautiful girls do? So the domestic CPU manufacturers secretly said, "I have to develop not only CPUs, but also software, supporting operating systems, etc."
The difficulty is equally great for the upstream and midstream.
A senior from Softmicro told me that the difficulty of manufacturing midstream integrated circuits is much more difficult than manufacturing atomic bombs and rockets. Well... I am ignorant...
Now let’s focus on the upstream and midstream.
Upstream: For example, semiconductor materials and the equipment required.
(For example, if you want to make a chip, you have nothing now, you definitely need raw materials and equipment to make the chip. Raw materials are only used once and then gone, but equipment needs to be depreciated...
For upstream raw materials, we divide them into three echelons based on the progress of localization.
(1) The first echelon: target materials, packaging substrates, CMP polishing materials, wet electronic chemicals, and some packaging materials. The technical standards of some products have reached the world's first-class level, and local production lines have achieved medium and large-scale supply.
(2) The second echelon: silicon wafers, electronic gases, compound semiconductors, and masks. The technical standards of some individual products have reached the world's first-class level, and local production lines have already supplied them in small batches or have great strategic significance, so there is a strong willingness for policy support;
(3) The third tier: photoresist. There is still a large gap between the technology and the world's first-class level, and mass production has basically not been achieved.
For the equipment required for upstream production.
For example, photolithography machines, etchers, thin film equipment, diffusion/ion implantation equipment, wet process equipment, process inspection, etc.
Upstream localization is difficult, whether it is semiconductor materials or equipment required in the semiconductor manufacturing process, all are difficulties.
However, the most difficult and urgent part for China at present is the midstream semiconductor manufacturing and preparation link.
Just like, I gave you pork and a cooking pan, but you can't make me a plate of fish-flavored pork shreds. Isn't this driving me crazy! ! ! ! Now even if the materials and equipment are localized.
Compared with the upstream, the biggest breakthrough point is the midstream. After the midstream is localized, the localization of the upstream will be put on the agenda immediately. However, if the midstream is not localized yet and we only focus on the upstream, will it be too "sell-research style"? (Of course, the upstream has room for growth in the long run)
Here we discuss the difficulties of midstream
The midstream uses equipment and raw materials to prepare semiconductors, including integrated circuit design, silicon wafer manufacturing, integrated circuit (IC) manufacturing, and integrated circuit packaging and testing.
A. Integrated Circuit Design
This step includes the front-end, middle-end and back-end:
The front end is to design the circuit (circuit diagram)
The middle section is the verification circuit, right?
The back end is to design the circuit layout (at the back end, the drawing is the actual picture)
At the end of the backend, you draw the layout, a bit like a fashion designer!!
Just kidding, that picture needs to be drawn layer by layer. It is actually a three-dimensional thing. Drawing the cross-section layer by layer... It is very difficult~
B. Silicon Wafer Preparation
Mining and purifying semiconductor materials. Including (crystal pulling - slicing - grinding - chamfering - etching - polishing - cleaning - testing)
C. Integrated Circuit Manufacturing
It is to form integrated circuits on the surface (companies that do this, such as SMIC)
This step requires the most money and is the most complicated. It is more difficult than making an atomic bomb.
The steps here are: oxidation-photolithography-etching-polishing-ion implantation-deposition-metallization-testing
D. Integrated Circuit Packaging and Testing
Testing is often carried out in the packaging factory, so packaging and testing are often treated as a whole, called the packaging and testing industry.
To be frank, the packaging and testing was the earliest breakthrough in our country and it is also the most mature.
So what are the difficulties in midstream? (Remember, midstream manufacturing includes the ABCD mentioned above)
The most important thing is that the IC manufacturing process is difficult. (It costs a lot of money) (It’s important to note that it’s not the IC design process in the middle (that’s A)!!!)
Why can’t we make a breakthrough in the integrated circuit manufacturing link?
The government built a manufacturing plant (that is, step C), spent a lot of money, and bought expensive equipment, but in the end it was left idle and became an empty factory. Why is this?
That's what we need talent for
Next, we will compare steps A and C in the middle stream.
First, let’s compare the development of disciplines:
A is integrated circuit design, and the schools with better development are Fudan University and Tsinghua University. However, in terms of design and manufacturing, my country's integrated circuit design, that is, step A, has developed relatively well. Although there is a gap with the United States, the gap is not as obvious as that of step C.
As for step C, the disciplines at Peking University are more focused on this area and are more cutting-edge (compared to other domestic universities).
Secondly, let’s compare from the perspective of capital investment:
A does not require much investment and has low production costs. More than a dozen people can roll up their sleeves and get started.
Step C, I'll go, not only build a factory, but also buy equipment (a photolithography machine bankrupted me), and the profit risk is high. Foreign countries beat me by a mile.
Let’s compare it from the perspective of employee salaries:
Step C needs to make a breakthrough the most, so it should spend a lot of money to recruit elites. But, hey, it is the capital investment in the previous step that everyone should know that the boss of step C has little money, so where can he get more money to give to employees? He can't even afford to pay for his own belt.
Step A, which is integrated circuit design, a Tsinghua student can earn 300,000 yuan a year after graduation. However, if he chooses step C, he can only earn 8,000 yuan a month.
Maybe everyone thinks, why do modern college students have to become money machines? Just go to work in C. But for students who work in C, patriotism is one thing, but if you want to recruit more talents to jump out of A and flow into C when they are employed, you also need compensation. Otherwise, the pressure of life is so great and the housing prices in Beijing are so expensive.
It’s like the salaries of internet celebrities beat scientists. Although, internet celebrities make money through their own efforts, and their live broadcasts also create value, at least making people happier and more entertained. However, if a country wants to improve its competitiveness, entertainment is not enough, it also needs to improve its hard power. These hard powers need money and policies to support them, just like this pneumonia, scientists who developed drugs should be given high honors and rewards. Otherwise, if everyone becomes an internet celebrity, what will they do when facing a strong enemy like the United States?
Finally, from the perspective of training model:
It is relatively easy to train a designer for step A. However, for step C, which is microelectronics, the equipment needed for research is extremely expensive. Which school can be so rich?
3.
“Localization” breakthrough
In order to solve the problem of localization, China has also set up some funds and issued relevant policies to promote the development of Chinese semiconductors and chips.
But from an intuitive point of view, compared to funds (the state will provide support, and some investment institutions have seized this opportunity to invest in leading companies in subdivided industries) (of course money is definitely needed, but I think money is the least valuable weapon in this regard), what our chips and semiconductors need more is talent.
On the one hand, from a technical point of view, we can poach people, especially experienced bigwigs.
On the other hand, the government needs to take certain measures to make the salary of workers in step C higher than that in step A, so as to attract many young graduates.
Finally, it will also fundamentally improve the talent training model for semiconductor-related industries, such as establishing corresponding research institutes. College students will be trained here to quickly connect their knowledge with reality, and after graduation, they will directly enter corporate factories to contribute to integrated circuits.
If the youth are strong, China will be strong. However, the entry point to solving these problems, such as how to build an education system to cultivate Chinese youth and how to enable China's aspiring young people to maximize their contributions, is precisely the entry point for the "localization" of chips and semiconductors.
As a finance major, I am ashamed to some extent when facing the "current situation of China's chips". Most finance students are quite impetuous, and everyone is competing for salaries, whether they have Beijing hukou, and investment bank internships.
Generally speaking, the salary of a finance graduate is higher than that of a semiconductor manufacturing employee. Actually, writing this article makes me a little sad. I wonder what would happen if many excellent talents were to work in semiconductor manufacturing instead of finance (after all, the supply of finance is now in excess of demand)? Would China be better off now? )
I don’t have relevant academic support to study the semiconductor industry and develop related products like a scientist to help the development of China’s semiconductor industry. The only thing I can do is to use my own industry research logic to analyze the upstream and downstream relationships of the industry, the competitive landscape and financial status of the target companies, and invest money in the most valuable companies to promote the development of China’s semiconductor companies and take “localization in China” a step further.
Although "localization in China" is a difficult hurdle, we will overcome it sooner or later!
about the author:
Hello everyone, I am Guo Xiaolan, a first-year graduate student at the School of Software and Microelectronics at Peking University. I have established the "Lala Internship Diary" with my good friend Wenrui, and we have 60,000 fans. We strive to become the cutest and most informative job search knowledge sharing platform on the Internet.
*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.
Today is the 2228th issue of content shared by "Semiconductor Industry Observer" for you, welcome to follow.
Recommended Reading
★ Is CMOS miniaturization over?
★ Power semiconductors will usher in a new era
★ What is the bottleneck of automotive semiconductors?
Semiconductor Industry Observation
" The first vertical media in semiconductor industry "
Real-time professional original depth
Scan the QR code , reply to the keywords below, and read more
“Core” Epidemic ｜AI｜TWS ｜ARM ｜ Storage ｜ CMOS｜Texas Instruments｜MEMS
Reply
Submit your article
and read "How to become a member of "Semiconductor Industry Observer""
Reply Search and you can easily find other articles that interest you!
京公网安备 11010802033920号