Lesson - 'don't expect others to sell you core technology'
Global Times: A lot of information shows that the start of Chinese computers is not too late. Is that right?
Ni Guangnan: When the Computing Institute of the Chinese Academy of Sciences was founded in the 1950s, only a few countries such as the United States and the Soviet Union were capable of doing large-scale computers in the world. At that time, our first large-scale computer for the electron tubes was modeled on the Soviet Union, plus a computer room and ventilation facilities. A computer is as big and spectacular as a basketball court.
However, when doing research in areas such as weather forecasting and petroleum exploration, our own computers are not enough, so we still have to use imported computers. This will be subject to strict scrutiny by exporters, and exports to your computer will not exceed us. The level of your own computer. Even if you buy it, the foreign country will send someone to watch it, and the requirement must be used for the specified purpose.
Now, our supercomputer performance indicators are number one in the world. This is what we have made ourselves. From this example we can know that do not expect others to sell you core technology.
Global Times: 'Ark 1' is the starting point of Chinese chips?
Ni Guangnan: There are many starting points for the Chinese chip design industry, 'Ark 1' is one of them. Earlier in 1988, we calculated the company, which was the predecessor of Lenovo, and started to make a special chip called ASIC (Application Specific Integrated Circuit). Sun Zuxi's research fellow took the young man to the Singapore Design Laboratory's design laboratory for research and development. In the five or six years, he borrowed other people's conditions. We later developed a total of five ASICs for use in Hanka, Chinese character printers, and micro-computers. In 2000, the State Council issued “Several Policies to Encourage the Development of the Software Industry and Integrated Circuit Industry,” commonly known as the 18th document, that stipulates that the taxation of the software industry should be reduced from 17% to 3%. The chip design industry should be considered as the software industry. Therefore, since then, the chip design industry in our country has rapidly developed. Chip design is similar to software, depends on human intelligence, has little to do with the industrial base, and we are not far from foreign countries.
On the other hand, in the field of chip manufacturing, there is a serious shortage of funds. No one is willing to invest capital in such a high-risk area where the talents are scarce, and the profit cycle is long and the probability of failure is relatively large. Therefore, chip manufacturing has not improved much. The level is very low, it can only be called 'try' and not talk about the industry.
Difficulties - 'Our software, who comes first?'
Global Times: What do you feel is the most difficult to break through the core technology of the chip industry?
Ni Guangnan: There is no doubt that the manufacturing process of the chip is our short board. It is somewhat similar to the traditional industry and depends on the industrial base. We often see on TV the scenes of precision manufacturing in a dust-free environment in a white coat. It is that we are behind many. Chip manufacturing is a capital-intensive, talent-intensive, technology-intensive industry that requires the nation to make up its mind and absorb social funds to make big investments to catch up.
At the same time, chip manufacturing is also special compared to traditional industries. Its economies of scale are particularly significant, requiring large investment. For example, SMEs in traditional manufacturing industries can survive by producing one or two parts, but only large companies can make chips. Go on.
Take BOE as an example, it belongs to a branch of chip manufacturing industry, and it only began to profit after 10 years of continuous investment. Huawei, BAT are not involved in chip manufacturing at present, even for such giants, the difficulty involved in chip manufacturing alone is still Too big, but they will participate in some form.
In addition, the construction of 'ecological' is also very important. The chip and operating system form the basis. It has a large number of hardware and software to form a system, and then develop a large number of applications to form a support system. This is an 'ecological'. Many large multinational companies None of the 'ecological' hurdles, such as the Wintel (Windows+Intel) system dominated the PC, but it was defeated by Apple and Android in the mobile space. The more no one used it, the less application support was. The less support is, the less vicious it is. This is a vicious circle. Once 'eco' is formed, it is a fait accompli. Monopoly is a kind of Matthew effect. The stronger the stronger, the monopolist's status can hardly be shaken.
Global Times: In the face of manufacturing and ecological 'two big mountains', where is China's chip industry going?
Ni Guangnan: It is necessary to increase core technology research and development and marketization guidance. The importance of research and development is needless to say. Why do we need to “marketize and guide”? For “ecological” construction, sometimes we can be compatible by purchasing commercial licenses. The more common 'ecology' in the world, but many times it is not for you, such as Wintel 'ecology' is not available, and fantasy Wintel is very easy to get rid of a '301 survey'.
Then another way is to have a market first, so that this 'ecology' gets a virtuous circle in the market: Some people use it, use the process to improve it, improve it to use more people, and then more people do application support to make it improve. But the problem that always exists on this road is: Who will be the first to use our software?
I think that China's government procurement market can be brought up. This is a market that we have not yet opened to the WTO. Our own software is used by the government itself. China has a large body, only government-purchased computers, and some 20 to 30 million are The market size is equivalent to one country in Europe. This advantage can be used to improve the 'ecological' construction.
Global Times: According to your assumptions, if China is determined to catch up, how long will it take to achieve world-class standards?
Ni Guangnan: As far as chip design is concerned, there is no big gap between us and the world's leading level, and the design does not require much capital investment. The monopoly is not very strong. As long as the planning is good, we can see it in about three to five years. Big effect. The chip manufacturing must be referenced according to the “Made in China 2025” plan. It will take at least 10 years, 8 years, or even longer. “Ecological” construction depends on our subjective initiative. Sort by difficulty , Manufacturing is the hardest to achieve catch-up, 'ecology' is the second, and design is the easiest.
Trends - 'The current boom is a good thing, but it should not be everywhere'
Global Times: Now that Chinese Internet companies are booming in the field of applications, it is said that taking away software and sharing bicycles will not be able to achieve great power status. Is this?
Ni Guangnan: There are many kinds of innovation, core technology innovation, business model innovation, management innovation, institutional innovation, etc. But I think core technology innovation is the foundation. Model innovation seems to be successful, but it needs the support of core technology behind it. So core technology innovation cannot be relaxed. I often talk to young entrepreneurs. They will think that the model innovation is faster, but I remind them that without core technology, you may be replaced quickly.
Global Times: The topic of 'China Chip' triggered an upsurge, and the speed of project financing and start-up has been significantly accelerated. What do you think of this phenomenon? Any suggestions?
Ni Guangnan: This is definitely a very good phenomenon. It shows that from the country to the enterprise, they all started to make a determination to break through. The big direction is right. But it should be noted that the power of all walks of life is a little scattered. According to experience, the chip industry needs If the funds are concentrated, if the development is decentralized, the final result may be that everyone is not strong enough. Therefore, I think we should better integrate and coordinate resources, not to blossom everywhere, and to disperse power.
In the international chip industry, monopoly oligopoly has long been formed, and small businesses in the future will have a hard time surviving and may eventually become useless. When it is necessary to break the monopoly, it cannot rely entirely on market laws, otherwise you will not be able to enter. The country originally had the advantage of 'focusing on the big things' and this advantage should be used. After all, compared to the international oligarchs of the chip industry, our investment is still too small.
Global Times: Why do we have to make our own chips and operating systems? Many countries in the world do not have these core technologies.
Ni Guangnan: It depends on what a country’s goal is. If it feels that it is enough to be a big country on the Internet, then we don’t have to pursue any core technology. But our goal is to build a strong country on the Internet, not to take the road to our core technology. May make you catch up.
In short, some things, if you think it is inevitable, then early determination is better than late determination, such as 'Beidou'. We decided to do it more than ten years ago. Now, Beidou is very useful. To engage in chip manufacturing, we may already have our own chips and won't be stuck.