SMIC’s Fourth Turning Point

This article is written based on public information and is for information exchange only and does not constitute any investment advice.

SMIC As the leading domestic semiconductor foundry, it is often regarded as a benchmark for A-share semiconductors and once topped the list of semiconductor market capitalization.

Since March 2021, due to the global chip panic and the vigorous domestic substitution, the A-share semiconductor market has experienced a 45-degree trend, and the Internet celebrity fund manager Cai Songsong has gone from a rookie to a god. However, in the hot semiconductor market, SMIC's stock price has been sluggish, underperforming the index by nearly 50% in the second quarter of 2021.

Every player who bought SMIC at the bottom because of its "cheap valuation" can only watch others eat meat while they themselves get beaten every day. Although SMIC, which was once obsessed with semiconductors, has also seen a rapid rise recently due to the extreme interpretation of semiconductors, its overall performance is still relatively backward. There must be something wrong when things are abnormal. Behind the sluggish stock price is a major change in the company's long-term development logic.

Figure 1: SMIC's stock price has been sluggish since the beginning of the year. Source: Wind

01

Once carried the dream of China's semiconductor industry

【1】The peak listing of the Science and Technology Innovation Board

On July 16, 2020, SMIC officially landed on the Science and Technology Innovation Board. Although it had been listed on the Hong Kong stock market before, SMIC was regarded as the first stock on the Science and Technology Innovation Board at that time. According to SMIC's initial plan for the Science and Technology Innovation Board, it planned to raise 20 billion yuan from the listing, but due to the enthusiastic subscription of investors, it finally completed the over-subscription and raised more than 53 billion yuan, which is 4 times that of China Railway Signal & Communication Corporation, the second-ranked company on the Science and Technology Innovation Board.

In addition to investment institutions, many well-known domestic semiconductor companies (such as Shanghai Xinyang, AMEC, Shanghai Xinsheng, Montage Technology, Zhonghuan Semiconductor, Will Semiconductor, Goodix Technology, ACM Semiconductor, Anjie Technology, Laimu Technology, Juchen Technology, Allwinner Technology, Pure Technology, Jiangfeng Electronics, etc.) also participated in the "red envelope distribution" link of SMIC's listing through the pre-IPO front line.

As for those who did not get the red envelopes, they naturally lined up in the group to give likes. It is no exaggeration to say that this was a feast that everyone enjoyed.

The good times didn't last long. After the short peak, there was a continuous adjustment. After July 2020, the Science and Technology Innovation Board entered a 9-month bear market, and SMIC was once considered the culprit. But in any case, SMIC, which received 50 billion, gained fame and fortune, and all it could see was a bright future.

Figure 2: Ranking of funds raised by the Science and Technology Innovation Board as of July 2021, Source: Wind

【2】Targeting TSMC

Friends who are familiar with the semiconductor industry chain may know that the foundry link is particularly critical in the entire industry chain. Once the foundry achieves a breakthrough, the explosion of chip design companies is only a matter of time. Otherwise, no matter how powerful domestic chip design companies are, they must look to overseas companies for advice.

After 20 years of hard work, by 2020 SMIC has overcome the 14nm process difficulties and has stepped into the road of advanced process. It is only two generations behind the world's advanced level. The only ones it has to catch up with are TSMC and Samsung.

SMIC itself made no secret of this in its prospectus: " In the field of logic processes, SMIC is the first foundry in mainland China to achieve mass production of 14nm FinFET, representing the most advanced level of independent research and development of integrated circuit manufacturing technology in mainland China; in the field of specialty processes, SMIC has successively launched mainland China's most advanced 24nm NAND, 40nm high-performance image sensors and other specialty processes, and cooperated with leading companies in various fields to achieve sustained growth in niche markets such as special memory and high-performance image sensors ."

It is no exaggeration to say that at that time, SMIC carried the hopes of the entire village.

Figure 3: Comparison of technology nodes of foundry companies, Source: Huatai Securities

For SMIC itself, the listing on the Science and Technology Innovation Board and the massive financing of 50 billion yuan are undoubtedly a declaration of its final charge towards advanced foundry processes, which can also be seen from the company's Science and Technology Innovation Board IPO fundraising projects. Among them, the 12-inch chip SN1 project will invest a total of 18 billion yuan, with the goal of raising the production technology level to 14 nanometers and below.

Figure 4: SMIC IPO fundraising projects, Source: Company announcement

Figure 5: Investment direction of SMIC IPO oversubscription funds, Source: Company announcement

The food and fodder have been prepared, and all that remains to be done is to march forward regardless of wind or rain.

03

When dreams suddenly become out of reach

【1】Huge amount of investment

As a seed player in mainland China, SMIC is undoubtedly on the right track and has very good strength. But no dream will succeed easily, first of all because the competition in semiconductor foundry has reached its limit.

According to IBS statistics, as the technology nodes continue to shrink, the equipment investment required for semiconductor foundry shows a nonlinear trend. Taking the 5-nanometer technology node as an example, its investment cost is as high as tens of billions of dollars, which is more than twice that of 14 nanometers and about four times that of 28 nanometers.

That is to say, even if SMIC raised 50 billion in the Science and Technology Innovation Board, it still does not have the funds to catch up with 5nm. In order to ensure its leading edge in the next generation of technology nodes, TSMC plans to invest another 100 billion US dollars in the next three years. The huge amount of investment has reached a staggering level.

Figure 6: Non-linear growth of foundry input, Source: Company prospectus

It is precisely because of this feature that the foundry link has become a battleground for a few manufacturers. At the 28nm node, there are more than 10 manufacturers in the world that can manufacture chips, but at the 5nm node, only TSMC is left.

For the pursuers, in addition to the initial massive investment that is difficult to overcome, the more difficult thing is that they may have to face long-term losses. Take R&D expenses as an example. SMIC has maintained about 20% for many years, while TSMC only needs less than 10%. In other words, the pursuers' large-scale investment is completed when they can basically not generate cash flow. This difficulty is an order of magnitude higher than the panel investment of that year.

This kind of dream that will not bear fruit for a long time is indeed something that ordinary people cannot persist in. However, as the only seedling in mainland China, we had reason to believe that with the support of the whole country, SMIC was likely to succeed. Take Samsung, the second largest foundry in the world, for example. Its foundry business started in 2005, five years later than SMIC, but it also succeeded in the end.

Figure 7: OEM elimination round, Source: New Era Securities

【2】Black Swan

Later, things did not develop as expected in the previous section, and the road to catching up, which should have been magnificent, was interrupted. With the well-known events that happened in December 2020, the core issue of SMIC's future instantly changed from "whether it can catch up with the advanced process" to "whether it can still use advanced process."

The conclusion is also heartbreaking. The ban is undoubtedly a direct blow to SMIC's dream of advanced process. As mentioned earlier, SMIC is only two generations behind TSMC in technology, about five years behind. For foundry companies, manufacturing equipment is the absolute core, and China's lithography equipment and other equipment are more than five generations behind overseas, and catching up is even more distant. Therefore, the inability to directly import semiconductor equipment containing American technology means that SMIC's dream of catching up with advanced technology has been directly frozen.

04

To the turning point

SMIC is forced to come to a turning point again. People are always used to magnifying short-term suffering. In fact, in the 20-year history of SMIC's entrepreneurship, it has experienced extremely bumpy experiences and has experienced several major turning points. In order to cope with future development, we first need to review the company's past ups and downs.

【1】Review of SMIC’s development history

(1) Turning point 1: Being surrounded by TSMC’s patents

In April 2000, SMIC was co-founded in Shanghai by Dr. Zhang Rujing and Academician Wang Yangyuan, who led more than 300 technical talents from Taiwan and more than 100 professionals from Europe, America, Japan, South Korea and other countries. SMIC was born with a silver spoon in its mouth. After its establishment, SMIC lived up to expectations and completed the construction of four 8-inch and one 12-inch production lines in the first three years. In 2004, it achieved profitability only four years after its establishment and was listed on the Hong Kong Stock Exchange and the New York Stock Exchange.

At that time, although wafer foundry was far from mainstream, IDM giants such as Intel and Texas Instruments were the industry leaders. However, TSMC, which later became the king, could no longer tolerate the threat from SMIC and twice took up patent weapons to sue SMIC in 2003 and 2006. In addition to paying more than US$300 million to reach a settlement, SMIC was more importantly forced to stop catching up with advanced processes.

In order to survive, SMIC chose to enter the DRAM market, which may be related to Zhang Rujing's background in DRAM at Texas Instruments.

In hindsight, this turnaround was undoubtedly a failure. Although the DRAM landscape was not determined at the beginning of the 20th century, after fierce competition, Samsung, Hynix and Micron became the three players who laughed at the end of DRAM. In 2007, when the global DRAM production capacity was seriously oversupplied, SMIC, which continued to lose money, also announced its withdrawal from the DRAM market.

Fortunately, the company did not give up the development of wafer foundry during this period and kept the spark.

(2) Turning point 2: Focus on mature technology

In 2009, Zhang Rujing resigned from SMIC, and the company experienced a struggle for management rights in 2011. With the return of veteran Qiu Ciyun, who left and then returned, to take over as CEO of SMIC in 2012, SMIC began to focus on mature processes and returned to the track of rapid development. It is worth mentioning that Qiu Ciyun also made great contributions to Huahong, the second largest mature process foundry in China, and now serves as a core executive of Shanghai Silicon Industry.

In 2012, SMIC turned a profit and continued to make profits thereafter. By 2017, the company's annual operating revenue exceeded US$3 billion for the first time, making it a second-tier wafer foundry.

(3) Turning point 3: Fighting again for advanced technology

Having recovered, SMIC has once again launched a charge towards advanced processes.

A representative event is that Liang Mengsong, known as the "chip madman", officially joined SMIC as co-CEO in 2017. The famous Liang Mengsong is the core figure who led TSMC and Samsung to become the leading foundry companies respectively. His joining SMIC undoubtedly indicates that the company is once again targeting TSMC.

As expected, SMIC made rapid progress within two years, achieving mass production of 14nm FinFET in 2019, directly breaking through the key threshold of 28nm, a mature process node. It has become the player with the most hope of advancing to the first echelon among the second-tier wafer foundry companies.

SMIC, which can continue to make profits with mature technology and rapidly make breakthroughs in advanced technology, has once again rekindled the market's confidence in it. It is not difficult to understand why SMIC, which was only prepared to raise 20 billion yuan at the beginning, was able to easily raise more than 50 billion yuan from the market.

SMIC is ready to buy EUV lithography machines, accumulate technology, and break through 7nm and 5nm step by step...

Figure 8: SMIC's mass production time for key technology nodes, Source: Company prospectus

【2】The Fourth Turning: How to Continue the Dream

With the occurrence of the black swan event, SMIC has undoubtedly reached another major turning point.

After facing extreme pressure at the end of last year, SMIC did not choose to lie low. The landmark event was on December 15, when SMIC announced the appointment of Chiang Shangyi as the company's vice chairman of the board, a second-class executive director, and a member of the strategic committee. Of course, there are many interpretations of Chiang Shangyi's appointment and Liang Mengsong's resignation. The focus of this article is not to sort out the love-hate relationship between Chiang and Liang, but to focus on SMIC's strategic transformation.

At this turning point, SMIC may choose to "return to mature technology". We observed three important events:

Even though dreams have been discounted, there is nothing shameful in adjusting to the dynamics of reality.

In the future, the global semiconductor foundry will reach a scale of 100 billion US dollars. In addition to the advanced processes that everyone is watching, mature processes will still occupy half of the market. After all, only a few high-end products such as CPU, GPU and mobile phone SOC can afford advanced processes. At present, among foundry companies, apart from TSMC and Samsung, other companies basically focus on mature processes, such as GlobalFoundries, UMC, Huahong, and Tower.

According to current industry tracking, the probability of mainland China successfully achieving full localization of 28nm in the near future is not low. Domestic chip design companies, except Huawei HiSilicon, rarely get involved in products within 14nm such as CPU, GPU and mobile phone SOC. Therefore, SMIC's transformation to mature process products is also the only choice to work together with domestic companies to break through the mature process chip market.

In addition, the current epic shortage in the semiconductor industry has caused the price of mature process chips produced on 8-inch wafer lines to soar, and the profitability of mature process foundries has also increased rapidly. Currently, SMIC, which has seven production lines, has turned around and started to work on mature processes again. In the era of shortages where production capacity is king, its chances of success are naturally better.

Figure 9: Global foundry industry market overview, Source: Company prospectus

05

Final Thoughts

Overall, the transition to a mature process is undoubtedly a life-saving move for SMIC. With no long-term dream in sight, the company's stock price has continued to fall. And for the transition to a mature process to be successful, the company must prove that its ROE can continue to improve, otherwise the market will choose to vote with its feet.

Of course, as a person who cares about the development of the motherland's semiconductor industry, we hope that this turning point of SMIC can become a transitional move like the "turning point 2" in its history, especially since Liang Mengsong was still retained after the resignation farce, which has left us with more thoughts about the spark.

Spreading out, this forced transformation of SMIC also reflects profoundly the technological lag of domestic semiconductor equipment companies, which has become the biggest constraint on the localization of semiconductors. SMIC has not supported excellent equipment companies in history, which has also laid the groundwork for today's predicament.

The capital market has also begun to realize this problem. Therefore, the stock prices of listed equipment companies such as China Microelectronics Corporation, North Huachuang, and Xinyuan Microelectronics have doubled rapidly recently. The baton of SMIC's dream that was defeated has also been passed to equipment companies. Whether the domestic semiconductor industry will play out the script of "mountains and rivers are so steep that there seems to be no way out, but there is a village with willows and flowers" in the future depends on the efforts of the new flag bearers.