Who is hyping up the “third generation semiconductors”?

At noon on June 17, a report from a foreign media triggered a frenzy in A-share semiconductor stocks. In response to this, Wang Zhong, a core person in a key semiconductor laboratory in China, told AI Finance that this wave of reports was taken out of context and was a form of hype.

According to the report, China has formulated a plan to promote the development of "third-generation semiconductors" and has prepared about $1 trillion in funds for the plan. Subsequently, self-media rushed to follow up.

On that day, the A-share semiconductor sector surged, with all 46 constituent stocks in the semiconductor industry index rising, and 38 constituent stocks rising by more than 5%. Even Cai Songsong, a star fund manager who bet on semiconductor stocks and a member of Noah Fund, returned to the hot search on Weibo, with the estimated net value of the fund he managed rising by 7.28% on that day.

Amid the enthusiastic response from the capital market, Wang Zhong told AI Finance that, as far as he knows, the background of this matter was actually a meeting held in Beijing on May 14. According to the information from the "Gongxin Toutiao" public account, at this meeting, the Ministry of Science and Technology reported on the preparation of the "14th Five-Year Plan" National Science and Technology Innovation Plan, and the leading group members and relevant department leaders discussed it. The meeting also discussed the potential disruptive technologies of integrated circuits for the post-Moore era.

"There is no explicit mention of third-generation semiconductors, and a trillion-dollar investment is out of the question. The global investment is not that large, and the global market is estimated to be only 10 billion US dollars, " said Wang Zhong.

He said that the current reports were vague, but they had touched the nerves of the market, heated up public opinion, and also hyped up semiconductor stocks. He expressed concern that this might mess up the domestic market layout.

Yu Kunshan, secretary general of the Third Generation Semiconductor Industry Technology Innovation Strategic Alliance, also told AI Finance that he did not know where the report came from. Liu Wei, a domestic silicon carbide academic, also told AI Finance: "The news was all over the place yesterday, and many people came out to refute the rumors."

So what are third-generation semiconductors? Why is there so much hype about them? Are they more advanced than the so-called first- and second-generation semiconductors?

The above-mentioned expert told AI Finance that, in fact, the third-generation semiconductors are not an upgrade of the first- and second-generation semiconductors. They coexist and each has its own application scenarios and advantages.

The first generation of semiconductors is based on silicon materials, which is also the most commonly discussed semiconductor concept among the public. It is widely used in mobile phones, computers, televisions and other fields. For example, Intel's CPU and Huawei's Kirin chip both use silicon-based semiconductor technology. The second generation of semiconductors is represented by gallium arsenide and indium antimonide, which are mainly used for power amplification and are used in satellite communications, mobile communications, navigation and other fields.

The third generation of semiconductors, represented by compound semiconductors such as gallium nitride and silicon carbide, are mainly aimed at three markets: optoelectronics, power electronics and microwave radio frequency. To put it more simply, mobile phone fast charging, new energy vehicles, rail transit and the national power grid are several major markets in the civilian field.

Wang Zhong further explained that the third-generation semiconductors are currently "still a niche market" and mainly serve as a supplementary market. 90% of today's semiconductor market is still first-generation semiconductors represented by silicon materials, and the second and third generations together account for only 10%.

The senior researcher told AI Finance that the term "third-generation semiconductors" is not rigorous. He has mentioned this on many occasions. The international term is wide bandgap, which refers to semiconductor materials with a bandgap width greater than 2.2eV. A person in the semiconductor industry also told AI Finance that relevant documents such as the Ministry of Industry and Information Technology and the 14th Five-Year Plan refer to them as "wide bandgap" semiconductors, not third-generation semiconductors. I don't know why foreign media are using this concept to make a fuss this time.

"The concept of 'third-generation semiconductors' was proposed in China mainly for cross-field exchanges and work reports, because industry terms are sometimes difficult to understand. But this term can easily give people the illusion that the third generation is better than the first generation," said Wang Zhong.

According to AI Finance, the world is basically on the same starting line for the hotly-hyped "third-generation semiconductors".

For a long time, the United States has used semiconductor technology to "choke" my country's science and technology field, which has also stimulated China's desire and action for self-reliance. However, the silicon-based semiconductor industry in China lags far behind, and it is very challenging to catch up. The silicon carbide and gallium nitride industries are believed to have opportunities to "change lanes and overtake".

Liu Wei, a domestic silicon carbide academic, told AI Finance that the gap between China and the United States in this technology is "just half a step behind, about two years."

"The gap in this field is relatively small, and China has the world's largest application market, so it may surpass foreign countries in the future." The above-mentioned person added, "I dare not say that we will surpass them in technology, but we will definitely surpass them in scale."

Many people said that the so-called "third-generation semiconductors" have a more autonomous and controllable industrial chain than silicon-based semiconductors. The core reason is that they do not require such a high-precision process. "Compared with the silicon-based semiconductor technology that uses 5 nanometers or even 3 nanometers, they are far from reaching the nanometer level and are all above 100 nanometers. Therefore, they do not need a particularly advanced lithography machine like ASML in the Netherlands, nor do they need a chip manufacturing plant like TSMC that invests tens of billions of yuan."

In addition, their requirements for EDA tools are also not that high. A silicon carbide expert told AI Finance that silicon carbide and gallium nitride semiconductors are still at the level of discrete devices and independent devices, not large-scale integrated circuits. "Although most domestic EDA software is still used abroad, it does not need to rely on so many complex functions. Therefore, the autonomy of EDA tools is not the main problem, at least not yet."

However, this does not mean that it has achieved full autonomy. Veteran Wang Zhong concluded that whether it is silicon carbide or gallium nitride, China has invested more money than the United States in the past 10 years, and there is no obvious generation gap between the overall technology and the United States, but in general, it is still subject to the underlying basic industries. "It's like Huawei can design the Kirin chip, but because domestic chip manufacturing is far behind, it will still be stuck." He said, "In fact, pulling out a carrot brings out a lot of mud, and in the end the problem still comes back to seemingly unrelated things."

"Take silicon carbide as an example. The core problem is the basic materials, such as high-purity carbon powder and silicon powder. We have a gap in purification technology. In silicon carbide equipment, we lack high-purity graphite crucible technology. When making devices, our lithography machines and photoresists are also a problem."

Liu Wei also told AI Finance that the manufacturing equipment for silicon carbide semiconductors still relies heavily on imports, especially expensive and entry-level equipment such as "epitaxial furnaces and ion implanters", which cost millions or even tens of millions of RMB, and are mainly imported from abroad. Gallium nitride also faces similar problems.

"There is still a lot of homework to do and a long way to go. Without any link in the supporting industry, nothing can be done," said Wang Zhong.

There are opportunities in China to seize this type of semiconductor technology, but it requires the systematic development of the entire system. AI Finance and Economics learned that Xidian University has a National Engineering Research Center for Wide Bandgap Semiconductors, China Resources Microelectronics' first 6-inch commercial silicon carbide wafer production line in China has officially entered mass production, Huachuang is developing core equipment epitaxial furnaces, and Shandong Tianyue and others are engaged in material research and development and production. In terms of application, Huawei, CRRC Group, BYD, etc. are engaged in the industrialization of related products. Previously, new energy vehicle Weilai also launched a gallium nitride mobile phone fast charging product to test related technologies... More and more companies have invested in it, which has promoted the development of the industry. But the more basic links still require the layout of industry-university-research cooperation.

Zha Yingjie, general manager of gallium nitride power chip company Navitas Semiconductor China, told AI Finance that he never believed in the concept of "overtaking on a curve". He believes that technology needs iteration and continuous investment. For technologies such as gallium nitride and silicon carbide, it seems that everyone has a similar starting point, but in fact, taking stock of the entire upstream and downstream core industries, domestic companies still need to continue to make efforts.

"Many things cannot be achieved simply by investing money. They require a lot of technical investment and constant trial and error. This really takes a very long time," said Zha Yingjie.

Although the construction of the entire industrial chain cannot be achieved overnight, industry insiders believe that the "third-generation semiconductors" are indeed a breakthrough for the country.

Veterans in the semiconductor industry believe that the "third generation semiconductors" are in the early stages of rapid development and are "catching up with foreign countries."

On the other hand, they can play an advantage in carbon neutrality and carbon peak. New energy vehicles, rail transit, energy Internet, these fields are all large areas where "third-generation semiconductors" play a role. "Now the market has also picked up, and people realize that these new technologies and products need to be used. This is an opportunity."

Previously, the fast charging chargers launched by mobile phone manufacturers once brought gallium nitride into the public eye. The reason is that compared with silicon-based semiconductors, gallium nitride technology has high power, fast charging, and small size, and can better meet users' charging needs.

More importantly, the investment required to develop "third-generation semiconductors" is much smaller than that of silicon-based semiconductors. Many industry insiders told AI Finance that the investment in silicon-based semiconductors is huge, and the construction of a manufacturing plant alone starts from tens of billions, while "third-generation semiconductors" can build a general-sized manufacturing plant with only 1 billion yuan.

However, Wang Zhong also warned that the focus on "third-generation semiconductors" should not develop into hype, so as not to repeat the mistakes of LEDs. Many years ago, gallium nitride made a major breakthrough in the LED lighting industry, and this news excited the entire lighting industry. The use of gallium nitride technology not only effectively solves the heat dissipation problem, but also increases the brightness per unit area by 10 times. Subsequently, the consumption of LED lights quickly became popular, but it also brought potential crises to the industry.

On the one hand, LED companies were caught in a price war. At that time, the price of an 18W LED tube dropped from hundreds of yuan to less than ten yuan, the market fell into chaos, and the profit margins of companies dropped sharply. On the other hand, listed companies used capital to attack competitors, and small and medium-sized enterprises that were eager to go public withdrew from the market due to lack of profits and broken capital chains.

Wang Zhong described that the entire LED industry has reached its darkest moment after being overheated. According to relevant survey data, by 2015, about 4,000 LED companies (including upstream chips and midstream and downstream lighting applications) had disappeared nationwide, especially downstream companies. Many local governments that invested in LED industrial parks have also accumulated a lot of debt.

He believes that after the concept of "third-generation semiconductors" came out, capital, self-media, and even some industry insiders participated in the hype to make a profit, which needs to be cautious. "We are very worried that the stock price will be hyped up again, and then it will fall down, which is a very bad thing." He said that external hype will make the industry impetuous, and the prices of talents and equipment will be higher. The disorderly increase of these costs will harm the development of the industry. For example, a piece of equipment that used to cost more than 2 million yuan has now been hyped up to 20 million yuan.

Regarding the fact that some second- and third-tier cities are now developing "third-generation semiconductors", Wang Zhong reminded: "Semiconductor investment requires funds, talents and management. Some places may be able to invest 1 billion, but where will the specialized talents and teams supporting process manufacturing and management come from? If these are not clear, it will be difficult to build the industry."

Overall, the "third-generation semiconductor" is a breakthrough, but it cannot be hyped indiscriminately, and the industry needs a rational atmosphere.

(Wang Zhong and Liu Wei are pseudonyms in this article)