The article further points out: "In the third quarter of last year, the first generation 14 nm FinFET process in this factory has been successfully mass-produced. After the planned production, SMIC South Factory will build two advanced integrated circuit production lines with a monthly production capacity of 35,000 pieces. 12 nanometer technology has also been introduced to customers, and the research and development of the next generation technology is also progressing steadily. The new production line will contribute to the future development of emerging applications such as 5G, Internet of Things and automotive electronics. "
Coincidentally, another domestic company with in-depth research on wafer foundry, Hua Hong Group, also disclosed at the supplier conference held recently that the company has made great progress on 14nm, and the research and development of the pilot process of more advanced technology nodes is also accelerating.
The announcement of these two leading domestic fabs marks that the foundry industry in China has entered a new stage.
Blue trail: the result of twenty-five years' pursuit
Since the establishment of Project 909, two fabs in Chinese mainland have been catching up with the leading manufacturers in the industry for 25 years. Look at TSMC in 1996. At that time, their revenue from processes below 1um reached 9.3%. By the year SMIC was established in 2000, TSMC's revenue had reached NT$166.2 billion, and its net profit had reached NT$ 65 10/00 billion, up by NT$ 127.3 respectively.
1996 to 2000 TSMC revenue ranking
As can be seen from the above data, even though the country has invested a lot of manpower and material resources, and even hired many experts from Taiwan Province Province and abroad, there is still a big gap between China's chip manufacturing industry and the world's leading level at that time. But later, Hua Hong Group (advanced technology is mainly promoted by Huali Microelectronics) and SMIC both grew rapidly in this pursuit, and the gap with the leader was never in sight. Now we can see the taillights of the leader. This is the result of many years of research by domestic chip manufacturing talents.
Take SMIC as an example, 20 10 was established in April, and construction started in August of that year. By September of the following year, SMIC had built three 8-inch wafer factories in Shanghai, setting a record for the fastest factory establishment in the world at that time. In September 2002, two 12 inch factories in SMIC Beijing started construction. In 2003, SMIC acquired Motorola's 8-inch chip factory in Tianjin.
Although SMIC walked faster in building factories, it was relatively slow in technology, which was partly related to some well-known reasons at that time.
According to relevant data, when the first factory of SMIC was still under construction, Zhang Rujing, the founder of SMIC, wanted to import production equipment with 0. 18 micron process from the United States. Even though it was not the most advanced technology in the United States (at that time, the technology of 0. 13 micron had been mass-produced), Zhang Rujing made great efforts to introduce these technologies. This situation continues to the processes of 0. 13 micron, 90 nm and 65 nm. Because it has kept its promise in the past, SMIC won the recognition of its partners and the US government at 45 nm.
But after 28 nautical miles, SMIC was "trapped" here again.
It is understood that SMIC has provided high dielectric constant metal gate (HKMG) and high KC process, including traditional polysilicon and gate finally. According to them, this is the technology they launched in the fourth quarter of 20 13. But in fact, for a long time, SMIC only provided 28nm polysilicon technology. Although the company said that the 28nm HKMG process was launched in the second quarter of 20 17, we can see from the reports of 20 18 and 1 in official website that before that, the yield of SMIC 28nm HKMG was only 40%, which is still a long way from being acceptable for mass production.
On the other hand, TSMC started the production of 28nm process at 20 1 1 because it has always had the concept of "absolutely leading in technology", and achieved rapid growth in the following years. The financial report shows that TSMC's 28 nm HKMG has contributed 27% of the company's revenue in the quarter when SMIC launched it. It is worth noting that TSMC's 10nm brought 1% revenue to the company in this quarter, which rose to 10% in the next quarter and soared to 19% in Q 18.
Revenue Distribution of TSMC Q2 20 17
As for 14nm, Liang Mengsong, co-CEO of SMIC, once said at the 20 19 Q2 financial report meeting that "the first-generation FinFET 14nm process of SMIC has entered the stage of customer verification, and the reliability and yield of products have been further improved".
Looking at Huali Microelectronics again, from the introduction of Mr. Shao Hua, Vice President of R&D in China District of 20 19 Semiconductor, we know that they have invested 8 billion US dollars in R&D by 20 10, and the company also has two factories in Zhangjiang and Cambridge. In particular, the second phase of Kangqiao Factory undertook the production tasks of advanced technologies such as 28nm to 14nm from Warwick. According to Shao Hua's statement at that time, Warwick was able to provide 28nmLP process and mass-produce HKC/HKC+ by the end of 20 19, and was also developing 22nm ULP and 14nm FinFET.
The news from the supplier conference in Hua Hong also showed that their 28nm processes were all mass-produced (including 28nm LP, 28nm HK and 28nm HKC+), and the research and development of 22nm made rapid progress, and 14nm also made great progress as mentioned at the beginning.
Only by laying a good foundation can they move forward more practically.
Internal troubles and foreign invasion: the necessity of further improvement
It is true that neither SMIC nor Huali Microelectronics can make further progress in technology in the future. Because with the miniaturization of the process, the technical difficulty is multiplying, and the cost to be invested is also huge. But considering the internal and external conditions, it is inevitable to develop advanced technology.
First, let's look at the external situation. In the past 20 19, the US government's actions against many China enterprises, including Huawei, broke the saying that technology has no borders. Many foreign media, including Nikkei Online, have also rumored that the United States will promote and hinder international leading fabs to serve Huawei and other China manufacturers. Although this statement has been denied by relevant parties, it is undeniable that it may become a "pawn" in the hands of American politicians.
One more thing, many internationally renowned media also categorically stated that the US government will restrict the supply of relevant manufacturers to domestic fabs, which will force the development of the domestic equipment industry. However, under the premise that foreign manufacturers are far ahead, it is difficult for some new equipment to cooperate with advanced fabs outside the mainland, such as TSMC. However, in order to make the equipment develop forward, if advanced technology is needed to cooperate with it, better results may be achieved. If this point can be well implemented in the end, it is bound to achieve a win-win situation.
Internally, on the one hand, as the recent news said, some domestic manufacturers, such as Huawei, have started to seek help from domestic manufacturers such as SMIC and Hua Hong, because they are afraid of the US "ban". Take Huawei as an example. In addition to the relatively backward technology, they need more advanced technologies such as 14nm, 7nm and 5nm. Coupled with the rise of big data, AI, 5G and other applications, more and higher performance chips are needed, and many domestic manufacturers are moving towards this goal. For them, if China has a reliable manufacturing process partner, it will definitely be the first choice for cooperation. But it also takes time.
Third, leading manufacturers Samsung and TSMC have taken a big step forward. Domestic manufacturers need to speed up their pace if they want to get the opportunity to compete with them on the same stage.
The latest news shows that TSMC's 5nm process has reached a yield of 50%, and the company also plans to promote the mass production of this process in Q2. Samsung has made a breakthrough in GAAFET, and plans to invest hundreds of billions of dollars in the next decade to compete with TSMC for the position of wafer foundry leader. These leading manufacturers also have research on advanced technology, EUV mask aligner and advanced materials in the future, which is their core competitiveness and is worth learning from domestic manufacturers.
However, for these two local manufacturers, in the future, each node will have to develop technology, or skip some nodes as needed and leap to a new stage. This is also a question to consider. Let's look forward to their next decade.