SMIC, sitting in the second position in the world!
In the first quarter of 2024, SMIC's financial data was announced, with revenue of US$1.75 billion (about RMB 12.644 billion) in the quarter, compared with US$1.462 billion in the same period last year, a year-on-year increase of 19.7%, and a quarter-on-quarter increase of 4.3% compared with the fourth quarter of last year. After more than a year of "U" shaped growth, ZTE International has finally come out of the trough and returned to a steady growth rate of revenue.
What's even more surprising is that this is the first time that SMIC's quarterly revenue has surpassed that of UMC and GF, two chip manufacturers, whose quarterly revenues are $1.71 billion and $1.549 billion, respectively. In other words, SMIC is the world's second largest wafer foundry, of course, there is still a huge gap with the first, TSMC's revenue this quarter is about $18.262 billion.
Revenue exceeded expectations and became the second largest in the world, and SMIC's stock price also rose.
However, if SMIC wants to continue to make great progress, there are still some problems that need to be solved.
The expansion of production capacity during the rapid growth rate has become the current high-pressure inventory, which affects its liquidity, cash pressure and profitability.
SMIC, which has come out of the trough, what should we do next?
It's getting better, but not completely
SMIC has come out of the trough.
In the first quarter of 2024, SMIC achieved revenue of $1.75 billion, up 4.3% sequentially. As can be seen from the sequential growth trend in the first quarter of 2021, SMIC has basically returned to a state of steady growth after experiencing a trough in 2022 and a recovery period in 2023.
However, SMIC still can't be happy too early on this revenue.
Wafer shipments in the quarter were 1.79 million pieces, an increase of 7% from the previous quarter, and the increase in shipments is a good thing, because the company's downstream shipments of some leading customer products have increased, which has increased the demand for wafers.
However, if you divide the revenue by the shipments, we can get that the revenue per chip in this quarter is about 975 US dollars per piece, while this figure is about 1002 US dollars per piece in the last quarter (a wafer can cut out 500 to 700 chips at the mobile phone SoC level), that is, the demand has picked up, but the price has not.
The sequential decline in selling prices will naturally affect the performance of gross profit margin, with SMIC's gross margin of 13.7% in the quarter down 2.7 percentage points from 16.4% in the previous quarter. Gross margin performance exceeded expectations, but it still declined sequentially.
Another problem is that SMIC's high inventory level has not been resolved when demand was weak due to the expansion scheduled during the rapid growth rate in previous years.
SMIC's inventory was almost 1.6 times its revenue in this quarter, and it can be seen from the trend chart in recent years that SMIC's inventory has been more than 1.5 times its revenue in almost every quarter since last year.
If you calculate the inventory turnover rate in this quarter, it is 0.55 times, which means that it will take almost two quarters for SMIC to sell the goods it has now. The problem is that enterprises are operating, and will not wait until all the inventory in hand is sold out before continuing to produce, if the market demand does not increase significantly, then more inventory will be piled up in the next quarter, and the cycle will start again, forming a high inventory pressure.
Don't underestimate the high level of inventory, which can cause a series of problems.
When a company is in a hurry to clear its inventory, it will naturally do two things, one is to reduce prices, and the other is to give customers more right to speak, such as the proportion of customer credit can be increased, the time for payment collection can be slowed down, and so on.
The former affects profitability, and we have analyzed that the decline in revenue and gross profit margin may be a side effect of high inventory pressure; The latter will affect the company's voice in the downstream.
Fortunately, from the perspective of accounts receivable, SMIC's voice in the downstream has not actually been affected.
On the contrary, SMIC's accounts receivable as a percentage of total revenue is actually on a downward trend. If compared with TSMC next door, TSMC's accounts receivable turnover in this quarter was 2.93 times, and SMIC's was 3.56 times.
Regardless of whether it is compared with itself or its peers, SMIC's downstream voice performance is good. This is a good demonstration of SMIC's position as an industry leader.
However, there is another hidden effect of high inventory, which is the impact on assets.
Inventory is counted in current assets, and the problem chip is a product with rapid technology iteration, but with the longer the inventory time and the technology update and iteration, this batch of inventory will be impaired, and it may be suffocated in the hand.
SMIC's current asset-liability ratio in this quarter is 1.63, and TSMC is 2.39 as a reference, which means that SMIC's money on hand and what can be exchanged for money immediately is 1.63 times its debt, while TSMC is 2.39 times. This value in itself is not a peace of mind, and inventories account for 21% of current assets, compared to TSMC's inventory to current assets ratio of only 10.9%. In other words, if SMIC's inventory can't be liquidated quickly, there will be pressure on the cash side.
This can be seen in the cash flow performance, the cash flow generated by SMIC's operating activities in the quarter accounted for about 28.3% of total revenue, compared to TSMC's 73.6%.
In addition to high inventories, SMIC also faced a problem this quarter, which was to increase revenue but not profits.
SMIC's net profit attributable to shareholders of listed companies in the quarter was 509 million yuan, down 68.9% year-on-year. The month-on-month decline was about 60%. If you break down the components of SMIC's profits, you will find that in addition to the decline in gross profit, amortization is also a major reason for diluting its profits.
As a wafer fab with heavy asset management (equipment, factory), SMIC's amortization expense is very substantial.
Last quarter, SMIC disclosed that the depreciation and amortization expense was 602 million US dollars (about 4.35 billion yuan) This quarter, ZTE did not disclose the depreciation and amortization expense separately, but disclosed the depreciation, amortization and EBIT profit was 6.295 billion yuan, and the income tax expense was 77 million yuan, which can be roughly estimated that the depreciation and amortization in this quarter is about 6.2 billion yuan, which is a large increase.
From the comparison of the three major operating expenses (sales, management, and R&D), the three major expenses in this quarter were 2.164 billion yuan, compared with 1.82 billion yuan last year and 3.1 billion yuan in the previous quarter.
Looking at SMIC's overall financial performance in the quarter, stable revenue growth is gratifying, but high inventories, reduced gross margin, and diluted net profit remain challenges ahead.
SMIC has improved, but it can't breathe a sigh of relief.
Expansion and expansion
Last quarter, SMIC said it plans to continue its 12-inch factory and capacity construction plans announced in recent years in 2024, and expects capital expenditures to be broadly flat compared to the previous year.
Judging from the level of capital expenditure in this quarter, the pace of expansion does not seem to have stopped - SMIC's capital expenditure in this quarter was 15.873 billion yuan, the same as the previous quarter and 1.83 times that of the same period last year.
As for the direction of capital expenditure, SMIC wrote in its annual report last year that "the capital expenditure in the reporting period was mainly used for capacity expansion. "In the past two years, SMIC's R&D investment has been expensed rather than capitalized, and if the situation has not changed much from this quarter and the end of last year, then it is reasonable to speculate that part of the increased capital expenditure in this quarter is used to expand production.
You must know that the decision to expand production last year was against the market.
SMIC's revenue in 2023 will be RMB45.25 billion, the first negative year-on-year revenue growth since 2020.
So, why does SMIC still have to expand production in the face of declining revenue?
The reason for this can be explained in its 2023 annual report.
According to the annual report, "from a geographical point of view, the regional development trend of semiconductors in recent years has been obvious year by year. The driving force for the localization of the semiconductor industry mainly comes from the scale of local market demand and the resilience of local economic development. From the perspective of Chinese mainland's industrial situation, as one of the world's largest semiconductor consumer markets, the mainland's integrated circuit industry is still dependent on imports to a certain extent at this stage. The scale of the existing integrated circuit industry in China, including the scale of wafer foundry capacity and process technology capabilities, still does not match the actual market demand. With the promotion of a new round of scientific and technological innovation, the domestic industrial chain has a large space for growth. ”
To put it in the vernacular, China's demand for semiconductors is very large, but now a large part of it is imported from abroad, and in the future, this piece will be localized and autonomous, so in summary, the four big words "the future can be expected".
This is certainly in line with the country's development strategy, so how to seize this future in the next period of time?
Let's take a look at two points, first look at the macro.
From the perspective of the downstream demand of the wafer industry, the most typical is consumer electronics, that is, the smartphones, smart homes, PCs, etc. that we often use, and consumer electronics will indeed show signs of gradual recovery from the second half of 2023.
Last year's Double 11, the sales of mobile phones, digital and computers, which were 194.6 billion yuan in consumer electronics, were the second and sixth largest categories among the top 10 products on Double 11. IDC data also counts that during the period from October 23 to November 3, the overall PC market sales increased by 1.4% year-on-year.
The revenue distribution of this quarter can also be seen that in the downstream demand, the smartphone business accounted for 31.2%, which continued to rebound, and was already the main force of the company's revenue growth in this quarter; However, the computer and tablet business accounted for 17.5%, and the industrial and automotive business accounted for a relatively low 7.2%, both of which declined to varying degrees.
Since the second half of last year, the increase in smartphone shipments, and the successive launch of new products by mobile phone brands such as Huawei and VIVO have slowed down the decline in mobile phone shipments to a certain extent. Moreover, since last year, under the trend of various All-in AI, various "large models" as gimmicks of intelligent hardware can add a fire to consumer electronics.
Looking at the micro level, SMIC's capacity utilization rate reached 80.8% this quarter, and SMIC's capacity utilization rate has been at a low level since the revenue showed a downward trend.
At this time, the continuous expansion of production will have two impacts.
The most obvious is that it will reduce the capacity utilization rate, and another problem is that don't forget that SMIC's inventory pressure is already quite large, and if the production capacity soars, the problems will be superimposed.
It seems that SMIC's expansion is more of a long-term strategic choice, and the short-term growth space is not very obvious.
Therefore, even if the future is promising, the expansion at this stage must withstand financial pressure. So SMIC came up with a trick to use the profits from investment to cover the expenses of operating assets.
For example, on March 26, SMIC issued an announcement saying that it sold all the shares of Changdian Technology it held, and 228 million outstanding shares were traded at a price of 29 yuan per share, so it almost cashed out more than 6.5 billion yuan.
Throughout 2023, SMIC's investment department has also received a lot of IPOs, including 13 invested companies such as Yutai Micro, Innosilicon, Tiancheng Technology, and Nanxin Technology, which have successfully landed in the capital market, and SMIC, as an early investor, will naturally benefit from it.
Kong Xin (pseudonym), a researcher in the chip semiconductor industry, said that SMIC's investment in the industrial chain is, on the one hand, to turn more upstream and downstream industries into "teammates", and on the other hand, it is to "transfuse" itself through the transfer of assets when the expansion investment is facing financial pressure.
When will the wind of AI blow
When is the future SMIC betting on?
In the wafer market in 2024, there are two main forces driving growth - AI and autonomous driving, TSMC is already full of oil on these two tracks, but if you want to eat these two cakes, the threshold is not low.
What the AI industry brings is the need for advanced packaging.
The so-called advanced packaging is to split and assemble or stack one or more DIIs (bare chips, which are cut from wafers according to the layout) with different functions to achieve the effect of integrating multiple chips. (Tiger Sniff)
Why do AI chips need advanced packaging? Because AI computing is a typical large-scale parallel computing, the complexity of a single algorithm is not large, but the number of calculations required is particularly large, which means that memory needs to be read and written extremely frequently, and advanced packaging can improve the memory and data transmission rate of the product, in other words, larger capacity and faster processing speed to keep up with higher requirements.
After the popularity of generative AI, the demand for advanced packaging has also increased. TSMC said at last year's shareholders' meeting that many orders for ChatGPT-related semiconductor products require advanced packaging, so the demand is far greater than the production capacity.
For example, TSMC launched the CoWos (Chip On Wafer On Substrate) process in 2012, but the cost of wafers using this process is 4000-6000 US dollars, while the cost of another advanced packaging process, InFo, is only 400 US dollars, so the process winner is one step closer to "winning the world".
Another problem is competition, it is not a secret that advanced packaging is the future trend, including Samsung, Intel and other foundries, as well as packaging and testing manufacturers such as ASE have long been deployed in advanced packaging technology. Samsung launched I-Cube, X-Cube and other processes; Intel's 3D packaging technology, Foveros chips, will be produced at Intel's newly upgraded Fab 9 chip factory in New Mexico, USA. (Observer.com)
In contrast, SMIC's advanced packaging layout is not ahead of its time. In fact, JCET has laid out advanced packaging technology by SMIC, and SMIC has also established a joint venture with it, but with the change of hands of JCET, how SMIC will further cross-border layout is also a question to be further observed.
Last year, many mobile phone manufacturers shouted "All-in AI", and then various "large model smartphones" and "AI mobile phones" began to be new, and AI chips, as the most important components of AI smartphones and other hardware, will also affect the performance of hardware, so how big and how long can the dividends brought by AI to the chip industry depend on whether head manufacturers such as Apple and Huawei can come up with killer AI applications.
Another thing that is seen as the main force of new demand is the "three-electric system" in new energy vehicles and the opportunities brought by autonomous driving.
The "three-electric system" in new energy vehicles needs power chips, so the growth point brought by this demand is mainly in the field of power semiconductors.
However, in recent years, almost all the increments of power semiconductors come from new energy vehicles, and the demand increment of photovoltaic and power industries has slowed down significantly, so all manufacturers are working hard for new energy vehicle factories.
In the most downstream market, there is a price war, and chip factories can't make money, either to reduce production or to reduce the cost of purchasing, no matter which one it is, it will affect the upstream wafer industry.
As for the demand for autonomous driving, it is more volatile.
On the one hand, there is the technology volume, TSMC once said that it has achieved the world's first in the production of system-level wafers (SoW) for Tesla. Tesla's Dojo wafer system processors are 5×5 arrays of processor chips that are placed on a wafer carrier and interconnected using TSMC's integrated fan-out (InFO) technology to enable wafer-level interconnects (InFO_SoW) (smart hyperparameters) so that Tesla's 25 chips can run like one big chip, and how to maintain the uniformity of these wafer processors is a test of the fab's technology.
In addition to technology, there is also a problem with domestic autonomous driving that it is not profitable.
Most automakers are "gifting" autonomous driving to car owners in the form of higher-level SKU purchases, rather than selling them separately like Tesla, and are ready to make a lot of money with software, so automakers will not invest heavily in this procurement. However, with the entry of FSD into China and the technological upgrading of major domestic car companies, whether it will usher in a more open policy is a potential positive.
Speaking of which, whether it is AI, new energy vehicles and autonomous driving, there is a cake, but when and who can eat it is still unknown.
Chip manufacturing is one of the most volatile industries in the world, without one. SMIC has said before that the company is in a "U" shaped growth, and this trough has passed, and the next question is how to win all the opponents.