On May 9, SMIC, a leading wafer foundry in the mainland, released its first-quarter results, with revenue reaching 12.59 billion yuan, surpassing UMC's 12.1 billion yuan and GF's 11.2 billion yuan, becoming the third largest wafer foundry after TSMC and Samsung.
This news quickly fermented on the Internet, causing heated discussions among netizens, and some netizens said that the gap between SMIC and TSMC is only EUV lithography machine, so is this really the case?
The feud between SMIC and TSMC
Many people don't understand that SMIC and TSMC are a pair of old enemies, and they once fought to the point of "you live and die".
In 1977, 29-year-old Zhang Rujing joined Texas Instruments, an American semiconductor company, at this time Zhang Zhongmou was already the vice president of Texas Instruments, and also served as the general manager of the semiconductor group.
However, the gap between the ranks of the two sides is too big, and there is not much intersection. In 1983, Zhang Zhongmou resigned from Texas Instruments, and then came to Taiwan to found TSMC.
It was not until 1990 that Zhang Rujing began to lead the team to participate in the management and construction of wafer fabs for Texas Instruments in Taiwan, Italy, Japan, Singapore and other places, and returned to Taiwan in 1997 to found Shida Semiconductor Co., Ltd. and served as its president.
Shida developed rapidly, and in just two years, it became a competitor of TSMC, and at this time, the "rival play" between Zhang Zhongmou and Zhang Rujing began.
Zhang Zhongmou, in collusion with the Lee Teng-hui authorities, secretly acquired most of the shares of the World Congress, and the shareholders of the World University readily agreed to the huge amount of $5 billion.
Zhang Rujing learned that "raw rice has been cooked into cooked rice", so he negotiated with Zhang Zhongmou and asked TSMC to build an advanced wafer factory in the mainland, Zhang Zhongmou agreed, but he did not take action for a long time, so if he delayed it, the daylily would be cold.
As a result, Zhang Rujing came to Shanghai to found SMIC, with more than 300 engineers from Shida Semiconductor and TSMC, and more than 100 experts from overseas to join.
With Zhang Rujing, the "king of factory construction", and a group of experienced scientists and engineers, SMIC has developed very fast. In just 13 months, SMIC mass-produced 8-inch wafers, setting a new record for the world's fastest chip factory to be built.
Just as SMIC was preparing to go public in the United States, TSMC began to make trouble.
In August 2003, TSMC sued SMIC in California for patent infringement and trade secret theft, demanding $1 billion in damages, which was the first time that wafer factories on both sides of the strait faced off in court.
SMIC's revenue at the time was only $360 million, and it simply could not afford such compensation.
As a result, TSMC filed similar lawsuits against SMIC in May 2004 and August 2006.
Six years later, on November 10, 2009, the exhausted SMIC was forced to announce a settlement in exchange for a total payment of $200 million to TSMC for a 10% stake in SMIC, and the resignation of its founder, Zhang Rujing, who has been banned from chip-related industries since 2010
Such a result can be described as a "fiasco", but Zhang Rujing said to his colleagues who came to see him off: "Sacrificing me in exchange for SMIC's way of life, this is good, we can't be beaten down, SMIC must continue to move forward bravely." ”
SMIC was lucky to survive, but missed nearly 10 years of golden development time, while TSMC has grown into a global wafer foundry leader.
The gap between the two sides is already huge
In terms of revenue:
In the first quarter of 2024, TSMC's revenue will be 132.455 billion yuan, and SMIC's revenue will reach 12.59 billion yuan, less than 1/10 of TSMC.
In terms of net profit, TSMC was 55.656 billion yuan in the first quarter, and SMIC was only 509 million, less than 1/100 of TSMC.
R&D investment:
In 2023, TSMC's R&D expenses will be NT$178.7 billion, or about RMB 39.8 billion, accounting for 8.3% of annual revenue.
SMIC's R&D expenses in 2023 will be US$707 million, or about RMB5.1 billion, accounting for 11.2% of annual revenue.
TSMC's R&D investment is significantly higher than that of SMIC, and the technological progress brought by it is naturally also leading.
Process:
TSMC announced the mass production of the 3nm process in December 2023, and has now started to develop the 2nm process.
Capacity:
TSMC sold 15.3 million wafers (12-inches) back in 2022, and its capacity now will only increase.
SMIC's total production capacity of 8-inch and 12-inch wafers in 2023 will be 6.074 million pieces, all of which will be converted into 12-inch wafers, with a production capacity of about 4.5 million pieces, less than 1/3 of TSMC.
SMIC's production capacity under construction is about 4 million pieces per year, which is still considerable, but TSMC has three wafer factories in Phoenix, USA, and new production capacity in Japan and Germany.
Equipment Ownership:
In September 2023, TSMC already has 80 EUV lithography machines, and it also plans to introduce 10 EUV lithography machines, according to this calculation, TSMC already has 90 EUV lithography machines.
SMIC's figure is 0.
TSMC has about 700 lithography machines, while SMIC has less than 200 lithography machines.
R&D personnel:
In the previous report, TSMC had 6,534 R&D personnel, and this number will grow to 8,000 with the completion of TSMC's global R&D center.
SMIC's R&D personnel in 2023 will be 2,362.
It can be seen that the gap between the two sides is all-round, not only in revenue and net profit, but also in R&D investment, R&D personnel, equipment quantity, process and so on.
With its own strength, SMIC may never be able to catch up with TSMC, but backed by China, the largest chip market, it can at least get infinitely close to TSMC.
In 2023, China's imported chips will be as high as 2.5 trillion yuan, excluding more than 600 billion memory chips, there will be more than 1 trillion left, and SMIC can win 30% of the share and catch up with TSMC in terms of revenue.
However, if you want to win these production capacity, it is not just as simple as throwing money, buying equipment, and building production lines, but also to realize technological upgrading, equipment, and material localization.
This involves the entire domestic chip industry chain, which requires the R&D strength of upstream chip design, EDA enterprises, the competitiveness of downstream core packaging and testing, and the help of semiconductor equipment and material suppliers.
In the upstream, we have excellent designers such as Huawei HiSilicon and Ali Pingtouge, but EDA is still a weak link, and the downstream JCET is still among the top in the world, but equipment and materials are our weak links.
Although domestic chips have made significant progress, they are really not optimistic about the current geopolitical situation, because we have to solve EDA, EUV lithography machines and dozens of semiconductor materials on our own, which seems to be challenging the "impossible" and creating miracles.
Whether this "miracle" can be created or not, in addition to the hard work of scientists and engineers, also requires each of us to support domestic chips, purchase domestic equipment, and understand national policies.
I'm a science and technology master, if you like it, give it a like!