China is making progress becoming less reliant on foreign supply of semiconductors, according to Chinese IC industry executives, but the country’s efforts still require setting specific intermediate goals and prioritizing those goals.

The guests from the Chinese semiconductor industry at the 2019 China IC Leader Summit roundtable discussed the unavoidable problems encountered in the development of China's IC industry — and offered advice from various perspectives — on how to proceed with developing a home-grown industry.



Recommended
China and it's Pursuit for Chip Self-Sufficiency (Part 1)


The Unavoidable IP Problem

The main battlefield is that ICs are still in high demand everywhere, including general-purpose processors and analog devices. The competition between Chinese and foreign manufacturers comes down to intellectual property (IP), and the unavoidable conclusion is that while China’s electronics industry has developed some IP, it is still lagging far behind its international competitors.

As a well-known IP supplier in China, Dai also said that there are actually some areas in which China’s IP is leading, such as video transcoding in the cloud and network chips. “There was a top US company that found us and said that we can customize their low-power chips. Why are they looking to us? Because our transcoding and coding have been the best in the world. Our chip has four times the original power. The power consumption is only 1/30th. After that, several large IoT companies in China asked for this IP," Dai said.

Wayne Dai

Wayne Dai, VeriSilicon CEO

IP is a Long-Term Effort

Dai said, "The Internet model is easy to learn, but doing IP requires a long-term struggle. It is not easy to cultivate in 8 or 10 years. We have seen some slow improvement, but it take its time. Why is the first word in ‘Science and Technology Innovation Board’ science? It is to focus on the long-term accumulation of hard technology, rather than the Internet business model."

Miao has worked in IP for 16 years. "I know how hard it is to evolve IP every year, how to focus on one thing, and how difficult it is to lead the industry," he said. “Like the MCU industry, most of the key technical performance issues are related to analog performance. Chinese companies lack the accumulation of knowledge in this area. But I see more and more people and companies joining the industry. There are a lot of applications in China. With the demand built into the system, IP may be one step ahead of others and gain a leading position in certain specific projects."

Michael E. Porter, in his book The Competitive Advantage of Nations, discussed the importance of industrial clusters. He observed that forming such clusters is the most efficient way to create self-sufficiency in a market. Miao also hopes that the investment in China's IC industry will help develop a Chinese cluster — make the market more concentrated and encourage the creation of a home-grown talent pool, which will promote the IC industry, including IP development.

Senhua Dong

Senhua Dong, deputy general manager of Empyrean Software Co., Ltd

Give Domestic EDA Vendors a Fair Chance

Outside of China, the stranglehold firms have on technology is not only in basic IP. At present, 95% of China's EDA market is controlled by foreign enterprises.

Empyrean Software is the leading EDA supplier based in China. Dong noted that the EDA industry is 60 years old, while Empyrean is only 10 years old this year. China has to acknowledge there’s a gap in experience.

Dong said, "To develop China's EDA, we must first have strategic positioning. For example, Empyrean 's full-market operation has hardly taken any support from the state in the past 10 years, because we firmly believe that the development of China's integrated circuits will definitely bring opportunities to EDA. One is to strengthen cooperation with domestic OEMs. For example, the cooperation between Empyrean and Huawei is very close. China's IC industry demand has driven China's own EDA innovation. This kind of industrial chain cooperation is very important.”

Xia believes that Chinese industry should not automatically support domestic EDA tools simply because they are from China. The Chinese factory must practice non-discriminatory sourcing. American companies suspect that Chinese buyers reflexively support the Chinese manufacturers, but that’s not what Chinese buyers believe. In many cases, Chinese companies are actually giving priority to foreign products because they are mature and easy to use.

It is important to develop domestically-produced tools, but the market requires that Chinese suppliers get no special support — local suppliers must compete fairly with foreign EDA tools. From the market perspective, the monopoly that foreign EDA suppliers have is not good for downstream customers. “Even for the employees of these international EDA companies, it’s hard for them to job-hop,” added Xia. “Any country should support new market entrants instead of using national boundaries to limit who can participate or not. This is good for the development of the market, customers, and the entire industry."

Developing a competitive local EDA industry is a long-term goal. Foreign manufacturers started 60 years ago and have developed through hundreds of mergers and acquisitions. It is unrealistic to expect China can become competitive in EDA in three years. Domestic EDA manufacturers must aim at 10 years and 20 years to do it.

Production Equipment

The situation for semiconductor manufacturing equipment is similar. Advanced equipment has been banned for sale to China. If the availability of production equipment remains a bargaining chip in global trade disputes, how is China to increase the localization rate of chips?

As a representative for equipment manufacturers, Cao acknowledge that China's semiconductor equipment is far from the world's advanced level, and that has become an important factor restricting China's semiconductor industry.

That said, Chinese companies are becoming competitive. Cao said, "When our equipment is good enough, they will remove it from the list of restricted exports. For example, the Bureau of Industry and Security of the United States Department of Commerce went to AMEC, and then went to SMIC the next day, where they saw our equipment running smoothly in SMIC’s fab. They removed the equipment from the embargo list six months later." Cao said, "The reason is that China has been able to produce enough equipment and good equipment, so it’s no longer meaningful to continue to limit security in the United States.”

Even though China has made great progress in production equipment after 10 years, the overall situation is still unsatisfactory. Several people have said before that the equipment can be bought without having to spend a lot of money on it. However, after the ZTE incident, the whole country has a unified understanding.

Cao believes that the key to localization of equipment is the collaboration of upstream and downstream of the industrial chain. Downstream, Chinese IC manufacturing enterprises have to be willing to use equipment made by domestic suppliers. "In fact, we also have this problem. We will often ask ourselves: will we be reluctant to use domestically-produced parts? When we ask manufacturers to use domestically-produced equipment, can we actively help domestic suppliers who provide parts and subsystems for us?" he said.

"Domestic companies will encounter this problem,” Cao continued. “I think my product is ready, but no customers want to help me, no customer can tell me the disadvantages of my products.” In the past two years, the government has done a lot of work to promote the integration between upstream and downstream, such as letting downstream companies lead the research on some equipment and parts. From the first day, you are tied to the downstream enterprises."

China_image8

Cao Liansheng, vice president of Advanced Micro-Fabrication Equipment Inc. (AMEC)

Talent, the Hottest Commodity

In the final analysis, the competition in the semiconductor industry is a competition of talent. How will China's semiconductor be able to build a strong pool of employees and assure a continuously high-quality stream of talent?

Huawei has always been regarded as one of the most attractive Chinese companies for talent. It is also the only domestic electronics company with salaries that are comparable to other international giants.

Xia believes that “Everyone may think that the talent gap is now large and will be better in the future. In fact, if the basic education of talent is not done well, the gap will always exist or even grow wider. The primary school classroom determines national strength. Now, primary school mathematics competition is subject to many restrictions, so how can the industry hope to surpass others?”

The second is the issue the treatment of talent. The United States often accuses China of subsidizing certain industries. It accuses China of offering market-based subsidies, rather than investing in research and development.

It puts Chinese companies in a bind. If they can use money to apply for a project or buy equipment, but not for pay competitive salaries, how are Chinese companies supposed to attract and keep talent? If China wants to improve the skills of scientific research personnel, it needs to help them on Hukuo (registered residency status), reduce taxes from education, provide medical care and housing, and treat them equally,” Xia continued.

Miao agreed. He believes that Huawei has become a benchmark in this respect, an example that other Chinese companies should learn from. In 2000, Xia was working at Cisco, and made much more money than employees of Huawei. Huawei was still able to attract talented engineers, however, and now it has surpassed other major international companies, demonstrating how important the incentive system is to the enterprise — let the good people get better treatment.

"We must also find ways to go abroad. If we cannot attract good foreign engineers to China, we can set up design centers abroad and hire talented local technical engineers in other countries," Miao said.

That said, there are excellent engineers in China, and they need to be supported, agreed Dai. Another part of the problem is that China is not producing enough top-rank engineers to fulfill its own needs. This creates a situation in which Chinese companies are competing with each other to attract top local talent. China’s Internet companies are making a lot of money, which means they can pay high wages. IC companies cannot always offer competitive salaries, however, which puts them at a disadvantage when trying to attract top talent.

Part of the issue is the lack of sufficient protections for IP. Companies and individuals want assurance they will be able to profit from their innovation, specifically so that they can afford to offer competitive salaries. This is why President Xi Jinping and Premier Li Keqiang favor strict protection for IP rights and are recommending severe punishments for violations of IP rights.

“This is not for the Americans,” said Dai. The implication is that Xi and Li are not making promises to mollify the U.S., which has long expressed displeasure about China’s approach to IP. The politicians’ desire to provide stronger protection for IP is legitimate.

Dai said, “It is to protect the investment in the country and in local R&D companies. Because if you don’t protect the bottom line, the whole industry will be very bad. In short, if your company makes money, you can pay the employees high wages. Please don't rely on the state to pay for the salaries."

Dong suggested that we think about talent issues from another angle. Everyone hopes that colleges and universities can train enough talent to fill this gap. Assuming that colleges and universities can train these students, will these students enter the IC industry? Even if students major in engineering, even in semiconductor engineering specifically, it is not certain that they will pursue careers in IC.

“When I was studying in 1997, the Institute of Microelectronics, Tsinghua University could enroll two classes a year, and almost graduated 60 students,” Dong recalled. “After 10 years, Tsinghua’s micro-houses can only provide one graduating class per year, of less than 30 students. At that time, 70% of the 30 graduates would enter the electronics industry. But that means one-third of those graduates have not entered the IC industry. Why?"

It might be the inability to offer competitive salaries. If companies can't offer the same salary as Internet and AI companies, they can't make it. Worse, the integrated circuit industry in China is located in several relatively concentrated areas such as Beijing, Shanghai, and Shenzhen — places where the cost of living is high.

"The only way is to make your business more profitable and to invest that money in research and development. That is the key to solving this problem." Dong said, "Huawei is now the only one that can challenge the Internet companies with competitive salaries, so they can attract more students and professionals to enter.”

China IC Leader Summit

A roundtable discussion, titled "Countdown: How Far is China from 40% IC Self-Sufficiency?" was held during the 2019 China IC Leader Summit. The event was organized by Aspencore's EETimes China, EDN China and ESM China Electronic Business