Stepping into advanced processes is key.

Recently, SMIC has made a lot of moves.

Buying equipment with a lot of money, transferring orders from Huawei HiSilicon, going to the Science and Technology Innovation Board, etc., walking on two legs, SMIC’s footsteps in attacking high-end manufacturing processes are fast.

You know, Huawei HiSilicon’s full line of high-end products have always been manufactured by TSMC, and it was extremely difficult for SMIC to intervene before. Recently, it has been reported that Huawei has handed over some 14nm orders to SMIC. Part of this is due to changes in the external environment, but the progress of SMIC’s technical capabilities in this high-end process is undeniable.

As a latecomer, SMIC, which has been catching up with TSMC and Samsung, can become a top student?

The stable and rapid development of SMIC

Founded in 2000, it ranked fifth in global revenue from zero to 2019, and SMIC has developed rapidly.

According to the data, shortly after its establishment, SMIC began to cooperate with the Belgian Microelectronics Research Center, Infineon, Elpida, Motorola and other major manufacturers to manufacture chips in the IoT field. After four years of development, SMIC has achieved dual listings on the New York Stock Exchange and Hong Kong. In the same year, SMIC’s 12-inch production line quickly completed several major links, including equipment installation, production line construction, trial production and commissioning.

It is worth mentioning that the completion of the construction, testing and production of the 12-inch production line in less than a year is not only a manifestation of its rapid development, but also indicates that one foot has entered the market. With the mature process, SMIC will open the way to enter the mainstream market, and its revenue growth will also begin.

Why so much emphasis on 12-inch wafers? In fact, the reason is very simple. The 12-inch process range covers 0.13um to 7nm, and its wafer size corresponds to a wide variety of product types, covering almost all device types such as logic, storage, sensors, CIS, RF and power semiconductors. Therefore, after SMIC enters the 12-inch threshold, the market it can enter is also very broad; at the same time, due to the improvement of the manufacturing process, the average cost is reduced, and its revenue performance will also be improved. This is the key to its next step.

Therefore, in the past 20 years, SMIC has almost maintained such a development rhythm. Looking back at the development process, its process technology is improving at a rhythm of every two years. In terms of 12-inch wafer foundry, it has moved from the bottom-level mature process to the top-level advanced process, achieving mass production of the 14nm process.

Especially in 2017, in order to break through the advanced process, SMIC dug up Liang Mengsong, a leader in the foundry field. After he joined the company, SMIC’s share price rose significantly, especially in October when Liang Mengsong joined the company, its Hong Kong stock price rose by 35.6%.

In order to seize the key development opportunities of domestic substitution in the past two years, SMIC has spared no effort in equipment procurement. For example, from March 12 last year to February 17 this year, it purchased related machines and equipment for the production of wafers from Fanlin Group at a cost of US$600 million (about 4.2 billion yuan).

It can be said that SMIC spares no expense in the selection of core managers or the procurement of basic equipment. As a leading enterprise in the domestic wafer foundry field, led by Zhang Rujing, Qiu Ciyun, Liang Mengsong and other bigwigs, SMIC’s 14nm process in more than ten years is still amazing enough, and its development can also be said to be Fast and steady.

The financial report hides hidden worries, and strives to improve the advanced process

With so much trouble, SMIC’s purpose is obvious: to be a top student.

But its road ahead is not easy. First of all, it needs to be clear that for SMIC, only by truly infiltrating the global high-end market, like Samsung and TSMC, can it enter the ranks of top students. To go global, on the one hand, it must weaken the impact of export controls on its revenue growth, which is an inevitable topic; on the other hand, building high-end production lines is a key constraint for SMIC to enter the global market. .

Let’s first look at the issue of dealing with export controls. SMIC’s return to the Science and Technology Innovation Board can be said to be the best move it has taken. Recently, after SMIC announced that it was preparing to list on the Science and Technology Innovation Board, some industry insiders said with emotion, “Finally, there is no need to be wronged in Hong Kong stocks.” The general reason is the same.

As early as a year ago, SMIC chose to withdraw from the US stock market for this reason, because withdrawing from the sector that is not conducive to itself can minimize the impact of changes in the external environment on the company’s development.

In fact, even in TSMC’s announcement, it is mentioned that politics will greatly affect the company’s revenue. Because of various factors such as technology patents, foundries will inevitably face the risk of market restrictions, and SMIC will inevitably face the risk of market restrictions. The practice of returning A shares is undoubtedly the best at the moment.

But despite the external influence, such influence is short-term and intermittent. Considering the development in the next decades or even a century, SMIC will still focus on technology research and development and its ability to penetrate the commercial market. .

You must know that although the global revenue ranks fifth, in terms of the market share of wafer foundries, TSMC, the industry leader, has won more than half of the market, and with its own hardware product sales, Samsung also accounts for With nearly 20% market share, GF, which has always been based on IoT chip foundry, has almost swept the IoT field, accounting for 8% of the entire market. Excluding UMC, SMIC’s market share of about 4% makes it so. Difficult to associate with “excellent”.

From the latest annual report released by SMIC in 2020, we can see that the single revenue structure and the lack of advanced process technology are all its shortcomings.

From the analysis of SMIC’s regional sales proportion, we can find that the proportion of domestic sales in its sales has been rising, reaching 60% in 2019. However, affected by the international environment, its sales capacity in North America has been greatly affected. weakened, and foundry business in other regions such as Europe did not grow significantly.

It can be said that in the domestic market, due to the advanced technology and the tide of domestic substitution, SMIC’s market share is unparalleled, and it can also support its continuous increase in revenue. However, over-reliance on domestic market demand, SMIC will not go far in the end.

And to look at it in detail, currently the regions that SMIC can radiate are mainly Europe, North America, China and the surrounding areas, which have a crossover with TSMC. Originally, TSMC has always dominated the pack with advanced manufacturing processes. kind of repression.

We can see that in terms of process technology, the main contributors to SMIC’s revenue are low-end processes such as 0.18 microns and 55 nanometers, and high-end processes with higher profit margins are rare. Therefore, the mass production of the 14nm process in the past two years has a milestone significance for the future development or revenue of SMIC.

Of course, the direct impact of market, competitive products and other factors on SMIC is that it is not easy to maintain revenue growth in the process of large investment.

As can be seen from the net profit statistics in the above figure, SMIC has only started to make money since 2012. Before that, it has been in a state of loss or no profit. And this year, there is news that SMIC will increase its expenditure from 1.1 billion US dollars to 4.3 billion US dollars, which is equivalent to putting all of last year’s revenue into it.

Although listing on the Science and Technology Innovation Board can meet its financing needs to the greatest extent and its value can be recognized, for SMIC, the key to becoming a top student lies in technological progress. Capacity and revenue won’t be a problem until profit margins are truly improved.

The number of patents has grown rapidly, and the market layout has a solid backing

Looking back, the path SMIC has taken, whether it is buying equipment at all costs, building factories, or building advanced manufacturing processes, is quite similar to that of TSMC in the past.

However, if it is only the accumulation of resources, TSMC cannot achieve its current status. Some industry insiders pointed out that Taiwan’s IC manufacturing led by TSMC can be successful, and the leadership of key talents such as Zhang Zhongmou and the investment of Taiwan Industrial Technology Research Institute play a key role, but from the perspective of historical development, the entire international division of labor has played a key role. The historical development trend of the pattern from IDM to the separation of design and manufacturing can be said to be the key driving force for it to surpass Intel and IBM.

If SMIC wants to become a top student, it also needs such an industrial opportunity.

Since the end of 2019, the sound of “the semiconductor industry is approaching an inflection point” has been heard endlessly. The key fuse is the failure of Moore’s Law, that is, the cost and performance advantages of the process technology are approaching the critical point.

This is a test for TSMC, and an opportunity for manufacturers such as Samsung and SMIC, which have been suppressed all the time. To put it simply, the shock of the entire semiconductor industry has given latecomers an opportunity.

SMIC’s grasp of the market has always been very precise. Whether it is withdrawing from the US stock market, returning to the STAR Market in a timely manner, or inviting Liang Mengsong, including investing in Changdian Technology, acquiring Italian LFoundry, launching new 12-inch and 8-inch production lines in Shanghai, Shenzhen and Tianjin to expand production capacity, etc., it is based on market changes. Immediate feedback.

But to maximize this opportunity, SMIC still seems to have more than enough power for dim sum.

In fact, it is not difficult to detect that SMIC is still just following, and has not shown too many ideas with its own characteristics. Therefore, we can say that it is very pragmatic, but we can not help but worry about its future development.

By analyzing SMIC’s current business, it can be seen that it is playing the card of “comprehensive”, that is, in terms of various process technologies, it is one of the few wafer generations in the world that can provide a complete process. factory. But in the advanced process, even if SMIC achieves 7nm, it is not easy to surpass TSMC.

At present, in order to cope with the failure of Moore’s Law, the leading manufacturers are not only fighting for 3nm or even lower processes, but TSMC and Samsung are also developing special processes such as advanced packaging. SMIC seems to have no time to care about these technologies related to the future market.

However, just like the rise and rapid development of any major factory, the overall trend of SMIC’s patents has continued to grow. According to statistics, in 2019, the number of SMIC invention patents has reached 644, while TSMC has only 340.

Recently, TSMC, Samsung, etc. are going to build factories in the United States. This battle for foundry construction is about to start, and the domestic foundry banner will undoubtedly be carried by SMIC. In any case, with the correct strategy, the growth of its patent numbers has become a key positive signal and has given us a lot of confidence.