Text | Value Institute
On Friday morning Beijing time, Hong Kong stock SMIC opened high and went high, and the morning gain once expanded to nearly 3%. Supporting the higher share price is the financial report for the fourth quarter of fiscal 2021 released the day before. Overall, SMIC's fourth-quarter revenue met expectations, profits, gross margins brought many surprises, and more importantly, its main mature process wafer production capacity was further expanded.
As the largest and most advanced wafer foundry in the Chinese mainland, the burden on SMIC's shoulders is heavier than ever in the context of domestic enterprises facing multiple constraints and extreme pressure from foreign giants. This excellent financial report has undoubtedly given the market a reassuring pill.
However, looking closely at SMIC's business layout and revenue structure, the problems are still obvious: insufficient supply chain control, brain drain problems still exist, equipment is too dependent on foreign giants...
In the case of the spread of the chip famine and the intensification of the internal roll of giants such as Samsung and TSMC, the future of SMIC is bound to be both opportunities and challenges.
Revenue reached a new high, and SMIC focused on capacity expansion
To some extent, SMIC and today's A-share "new stock king" Ningde era have many similarities: the industry is a head manufacturer that no one knows and no one knows, but people outside the industry only feel "unclear" and do not know too much about its business.
The Value Institute believes that for friends who are not familiar with SMIC, it is enough to remember only a few key points.
First of all, according to the current market share and performance, SMIC is the first wafer foundry enterprise in mainland China and the top five in the world, with a global market share of more than 5%.
Secondly, the chip wafer foundry business related to communication equipment and consumer electronics is SMIC's most important revenue pillar, and its products are widely used in logic chips, mixed-signal/RF chips, high-voltage-resistant chips and system chips. As the most comprehensive wafer foundry in mainland China, SMIC has customers in smartphones, smart homes, automobiles and other industries, and Huawei HiSilicon is an important customer.
As we all know, due to the tight supply chain and the spread of chip shortage around the world, the head semiconductor companies can be said to have made a lot of money in the past year. Judging from the data disclosed in the financial report, SMIC has performed well in key data such as revenue, gross profit, and core business capacity, which objectively led to a rise in stock prices.
On the one hand, SMIC's overall revenue and gross profit margin have exceeded market expectations, and in the context of growing demand for semiconductors, it has eaten wave after wave of dividends.
SMIC's sales revenue was $1.580 billion in the fourth quarter, up 11.6 percent from $1.415 billion in the third quarter and up from $981 million in the fourth quarter of 2020, up 61.1 percent, the data showed. The performance of gross margin was even more surprising, reaching 35%, almost doubled from 18% in the same period last year.
In the view of the Value Research Institute, the increase in gross profit margin can be mainly attributed to the price increase brought about by the shortage of chip production capacity.
Looking at the records, it can be found that since the second quarter of last year, head wafer foundries including SMIC, TSMC, Samsung, and UMC have made price increase decisions. SMIC announced a full-line price increase in April last year, with orders that have been launched remaining unchanged, and orders that have been placed and not online will be traded at the new price.
According to the calculation of the price increase time, most of SMIC's orders in the second half of last year have been executed according to the new price, which objectively pushes up the gross profit margin. According to the data, SMIC's single-chip wafer gross profit in the fourth quarter was $321, an increase of $49 sequentially, while the revenue of single-chip wafers increased by $94 from the third quarter.
(Image from SMIC's earnings report)
On the other hand, the continuous expansion of production capacity and the increase in capacity utilization rate show that SMIC still has the trust of customers, and the price increase will not have a negative impact on its performance in a short period of time.
According to the data, SMIC's capacity utilization rate in the fourth quarter was 99.4%, although it failed to achieve three consecutive quarters of breaking 100, it also maintained a very high level. From the shipment data, the equivalent 8-inch wafer reached 1723 thousand pieces, which is a big improvement over the same period in 2020.
In addition to the 8 inches, SMIC's 12-inch wafer production capacity is also expanding. What needs to be known is that SMIC is currently the company with the highest scale of mass production of 12-inch wafers in mainland China. In the case of UMC's pace of TSMC and Samsung, which has begun to stumble on advanced processes, and the trend of localization substitution is prevalent, there may be more 12-inch and 8-inch wafer supply vacancies in the market, which will also bring opportunities to SMIC.
It is worth mentioning that in the past year, SMIC has been actively preparing for capacity expansion plans, including investing nearly $4 billion to expand 12-inch and 8-inch wafer production lines. In addition, given the impact of entity list sanctions, SMIC has also stepped up the development of the FinFTE multi-platform.
Driven by many factors such as lack of cores, price increases, and expansion, SMIC's revenue, profit, and gross profit margin have all achieved substantial growth, which is certainly worth celebrating. But at the same time, we still have to face up to the gap between it and the three major foundry giants in the first echelon.
According to IC Insights, SMIC has been at the top of the Chinese mainland pure wafer foundry market sales rankings for the past few years, and is also firmly in the top five in the world, on a par with UMC, but significantly behind TSMC, Samsung and GROMFOUNDRIES.
Returning to the development process of the past few years, what is restricting SMIC to go further? Is it technology, talent or money? As the hegemon of wafer foundry in the mainland region, how far is SMIC from TSMC and Samsung?
This may be a common question in the minds of many people.
Growing pains: talent, technology, supply chain
Dating back to the Csrcical Social Responsibility Report released by SMIC before its listing, there is a set of data that deserves our attention: the turnover rate of employees under the age of 30 is serious, accounting for nearly 80% of all employees leaving.
The Institute of Value said in the previous preview of the semiconductor industry at the end of 2022 that the competition for high-end talents in the industry will be very fierce in the next year, and wafer foundry giants such as TSMC, UMC, and Samsung are using high-paying and thick jobs to attract talents. At this juncture, SMIC's serious brain drain rate and insufficient attractiveness to high-end talents have been further amplified.
The reasons for this phenomenon are manifold.
On the one hand, SMIC's salary expenditure level is not prominent in the industry, and it does not have much advantage over competitors in the same echelon, and salary expenses have even shown a downward trend in the past few quarters.
In particular, in the second quarter of last year, the number of SMIC R&D personnel was 1,785, and the total compensation expenditure was 230 million, with an average compensation of 129,000, compared with 2,419, 326 million and 135,000 in the same period of 2020. In contrast, TSMC carried out a large-scale structural salary increase in January last year, and most of the core technical personnel's salaries have increased by at least 20%.
In the second half of the year, the situation did not improve much. In the fourth quarter, SMIC's research and development expenses were $172 million, up 2.8 percent from the third quarter and down 11.5 percent from the fourth quarter of 2020. Judging from the trend of the size of the R&D team and the R&D expense ratio, SMIC's R&D investment has not improved much after the listing.
On the other hand, in addition to the lack of money, SMIC also seems to have deficiencies in management and corporate culture.
InfoQ has previously reported that some insiders broke the news that after the implementation of the dual CEO system in 2017, with Zhao Haijun and Liang Mengsong at the helm, the above problems became more prominent. According to data from China Times statistics, as of the first half of last year, SMIC's turnover rate was about 1.3 times that of its peers, and the proportion of R&D personnel fell from nearly 18% before the listing in 2019 to less than 14% at present.
It is worth mentioning that in July last year, SMIC announced the official departure of Wu Jingang, vice president of technology research and development — you know, less than two weeks before leaving, Wu Jin was just granted 160,000 restricted shares at the extraordinary shareholders' meeting. According to the China Times, after the implementation of the double CEO system in 2017, Wu Jingang had a retreat.
It can be seen that in addition to the salary package is not as good as TSMC and UMC, SMIC has some deficiencies in corporate culture and internal management, and it cannot provide similar growth space to TSMC and Samsung in terms of technology, which is an important reason for reducing its talent attractiveness.
However, the inability to retain talent is an old problem that has plagued SMIC for many years, but it is not the only problem: the supply chain is too dependent on foreign giants, and the gap between technology and TSMC, Samsung, and Grofounds has been delayed, which is restricting its development.
In terms of supply chain, it is an open secret that the proportion of SMIC's domestic equipment is low and key technologies are stuck by foreign giants.
Taking lithography machines as an example, SMIC has publicly complained more than once about the impact of ASML delays in delivery on its production capacity. In the second quarter of last year's earnings call, CEO Zhao Haijun said that SMIC's 28nm process wafer expansion plan has been delayed due to problems with equipment delivery.
Although Zhao Haijun did not directly name ASML, the information reviewed by the Value Research Institute shows that ASML said in the second quarter of last year that the delivery time of the DUV lithography machine order signed with SMIC in 2018 will be extended to the end of December 2021. In view of this, even if SMIC's DUV lithography machine is sufficient to support the operation of the existing production line, ASML's delayed delivery will undoubtedly adversely affect its capacity expansion plans.
In terms of technical technology, SMIC is lagging behind Samsung and TSMC at this stage, which is also the consensus of the industry.
Wafer products of different sizes can be divided into two sections: the former is a 28nm and above process, mainly used in low-end mobile phones, digital television, Bluetooth chips, PCs, wearable devices and other fields; the latter process is 22nm and below, mainly used in high-end mobile phone APs, baseband, CPU, graphics card GPU and mining machine ASIC and other high-end products.
At present, TSMC and Samsung have carried out crazy inner winding on advanced processes to bring wafer accuracy into the 7/5nm era, while SMIC is still most capable of 14nm and 28nm process wafers.
However, from the perspective of the development strategy of the past year, SMIC no longer seems to be stuck in advanced processes, but has instead re-invested its energy and resources in the expansion of the production capacity of mature processes.
The technical barriers are extremely high, and the future development trend is also very clear in the semiconductor industry, and at first glance, there is no so-called cornering overtaking opportunity. But for SMIC, building on its strengths and avoiding its weaknesses and dodging Samsung and TSMC may also be a strategy worth looking forward to.
Focusing on mature craftsmanship, SMIC does not have to go to black in one way
According to the data, SMIC's process wafer production capacity of 28nm and below rose to 18.6% in the fourth quarter, and the growth momentum is quite good. But overall, the proportion of 12-inch wafers representing mature processes (corresponding to sizes of 90nm and below) is still the highest, more than 60%, and revenue also reached $894 million, recording a substantial increase of 78.6% year-on-year. From the perspective of subdivision size, the revenue of 150/180nm process wafers accounted for the highest proportion of 28.6%; the revenue of 55/65nm and 40/45nm process wafers also reached 26.8% and 15.3%, respectively.
(Image from Longbridge Dolphin Research)
In the view of the Institute of Value, 28nm and below process wafers are certainly the focus of future competition for wafer foundry giants, and Samsung and TSMC have also shifted their energy to this battlefield early. However, from the perspective of demand, mature processes still occupy more than 50% of the market share, and their importance cannot be ignored.
As mentioned earlier, the application areas of advanced process wafers are mainly high-end mobile phone APs, baseband, CPUs and mining machine ASICs. However, in the downstream of the mature process, there is no shortage of Internet of Things, automotive MCUs, low-end mobile phone Bluetooth chips, wearable devices, base stations and portable electronic devices and other demand sides that have also grown rapidly in recent years, and the growth potential is definitely worth looking forward to.
In fact, in addition to Samsung and TSMC, the revenue pillar of the second echelon companies such as UMC, GROFOUNDRIES, and Huahong Semiconductor in recent years is still mature process wafers, and it is also full of money.
Taking UMC, which is the closest to SMIC, as an example, its main revenue in the first three quarters of last year came from 12-inch mature process wafer products. Among them, the proportion of 40/65nm and 65/90nm process wafer revenue has further increased, and the production capacity of 12-inch mature process wafers has increased by nearly 6% month-on-month, and has recorded quarter-on-quarter growth for 10 consecutive quarters.
For SMIC, what we need to do now may be to learn from the previous lessons and make up for its shortcomings in talent retention and supply chain management.
In fact, SMIC has made a lot of efforts in the past year to address these two shortcomings. For example, in the supply chain, SMIC enriches the supply chain team by investing in more upstream and downstream enterprises and adjusting procurement strategies.
In October last year, Zhang Xin, vice president of SMIC, said that the company's procurement strategy was officially changed from a low price to a 30% price reduction to supporting strategic suppliers through bulk procurement. In addition, SMIC will also establish a training base for industry, education and research talents in universities and research institutes within the United Nations to improve the replacement rate of domestic equipment and alleviate the dilemma of being stuck by foreign giants such as ASML.
Of course, SMIC will not give up the competition for advanced processes. Last September, SMIC announced that it would build another fab in Lingang, Shanghai, and planned to mass-produce 100,000 12-inch wafers per month after completion – the third time in the past year that SMIC has announced capacity expansion. And at about the same time, TSMC and UMC also announced the expansion of process wafer production capacity below 28nm.
According to media reports, TSMC has begun to deploy a more mature FinFET architecture last year to achieve more accurate cross-channel current control, challenging the 3nm process. If all goes well, TSMC is likely to achieve mass production of 3nm process wafers by the end of this year, and Apple will become its first customers.
The battle for advanced processes by the head wafer foundry giants may have just reached its climax.
Write at the end
In the memory of the Institute of Value, the grand scene when SMIC landed on the Science and Technology Innovation Board is still vividly remembered:
Plans to raise 20 billion yuan, but because investors are too hot to finally complete the over-funding of 53 billion yuan, four times the China Pass number, which ranks second on the science and technology innovation board; SMIC's super IPO also induced a collective carnival of domestic semiconductor companies, Shanghai Xinyang, Zhongwei Company, Weier Shares, Huiding Technology, Shengmei Semiconductor all participated in this event through pre-IPO...
For SMIC, the brilliance of the listing does not mean a long-term smooth ride, and its development path is not smooth. Along the way, it has carried a lot of praise, and the doubts it has encountered are not inferior to the vast majority of its peers. However, one thing is undeniable: SMIC's potential and upper limit are definitely the most anticipated among the domestic chip semiconductor companies.
For now, there is undoubtedly a big gap between SMIC and top giants such as TSMC and Samsung. From a rigid advanced process to a return to a mature process, SMIC has also undergone several strategic adjustments and is still trying to adapt to changes in the market.
But given the broad prospects of the chip semiconductor industry, we believe that SMIC's future is endless.