Domestic Semiconductors Meet Development Opportunities Under Lack Of Core

The lack of core market is still continuing, and the domestic market is in a relatively stable production and development environment, and the ability in some links is indeed in a dominant position, which undoubtedly becomes one of the important fulcrums to undertake more semiconductor external demand in this period of time.

The performance forecast is one side. Tongfu micro power Co., Ltd. and Xinyuan micro Equipment Co., Ltd. are expected to achieve a net profit growth rate of more than doubling in the first half of this year.

Accelerating production increase is the other side. According to semi statistics, global semiconductor manufacturers will start building 19 new high-capacity wafer plants by the end of this year, and start building another 10 in 2022 to meet market demand. Among them, mainland China and Taiwan will take the lead in the construction of new Fabs in the two years, with 8 new Fabs each.

However, the semiconductor industry itself is also a cyclical industry. Industry insiders told 21st century economic reporter that due to the characteristics of this industry, structural overcapacity will be inevitable at some stage in the future.

To face the continuous changes in the industry, the core is to increase domestic production capacity, but also seize the opportunity to enhance technical strength, the current is the best time.

Traceable missing core

It is not too much to describe the current chip shortage situation with "once in a decade".

Under the influence of external environment, the manufacturers of the industrial chain reserve inventory ahead of time, which to some extent led to the demand for chips from the second half of last year, which greatly exceeded the capacity available at that time.

At the same time, as 2019 is the downward development cycle of semiconductors in history, and considering the impact of the new crown pneumonia epidemic, many inter industry companies have not correspondingly increased investment in early 2020 or even before.

"2017 is a peak. Semiconductor companies often make a lot of investment during the peak period, and two years after the investment output, supply exceeds demand. Therefore, from the perspective of the whole investment cycle, the current increase in production capacity in 2021 is actually the result of investment in the first two years. " Sheng Linghai, vice president of Gartner research, told reporters including the 21st century economic report.

As the production capacity did not increase in time, the demand for 5g mobile phones and bitcoin, which was hot in the first half of the year but has now declined, and the additional demand for laptops, servers and data centers boosted by the epidemic situation can not be met“ We expect this time cycle to be delayed to the second quarter of 2022, which may be adjusted in the future. " He continued.

According to the current situation of increasing production capacity, Sheng Linghai believes that 55nm / 65nm will be a process with great growth from present to 2025. The reason is that the market demand will still increase greatly now and even in the next few years. In addition, 28nm, 14nm, 16nm have greater opportunities to increase.

"Most of the traditional processes are focused on 8-inch wafers, but at present, 8-inch wafers are in short supply because of the excessive capacity of 8-inch wafers in the past many years, which has led to price drops. During the trough period, many factories, especially some related enterprises in Japan, have closed the 8-inch production line He pointed out. At the same time, the demand of 5g mobile phones for PMIC (power management chip) and analog circuit has increased greatly. In particular, the manufacturing process of PMIC is concentrated at 180 / 150 nm, which mainly needs 8-inch wafer and a small part of 12 inch wafer support. The increase in demand leads to a serious shortage of related devices.

At present, there is no new plant investment for the 8-inch production line, and most of the investment is to expand production. However, to solve the shortage of 8-inch process completely, it is still necessary to shift 8-inch production capacity to 12-inch production capacity. Because of the large production capacity of 12 inches, its output can reach more than twice of the former under the same time conditions.

But the move to 12 inch wafers is not easy, which is why there are not many participants yet.

Sheng Linghai said that previously, there were various processes in 8 inches and they were very mature. If we want to redesign a 12 inch wafer, wafer thickness is a key consideration.

"At present, the OEM is following up the transfer, but it is difficult for IDM manufacturers to follow up." He added that because the number of analog power chips that can be designed on 12 inch wafers is very large, if no giant companies can digest them at the market level, they will inevitably face the problem of losing money in the future.

From the demand side, as the demand of mobile phone industry has not reached the ideal value, the shortage of mobile phone industry has been improved; Due to the low demand of consumers, the shortage of TV sets has also improved.

"We estimate that the overall market will be in the peak season in the second half of this year. As the second quarter is often the traditional off-season, the current new capacity will increase in the second quarter of next year, so the shortage situation is expected to ease. But in contrast to power chips, resource constraints will continue for some time, as it involves investment in 8 inches and the process of moving to 12 inches, which is expected to continue into the second half of next year. " He continued.

At the same time, it can not be ignored that the current increased investment will face structural overcapacity in the future, which will not be transferred by will.

Sheng Linghai believes that overall, current investment will lead to oversupply in 2023 and 2024. However, even if supply exceeds demand, some enterprises with better planning in advance, such as TSMC, will remain competitive in the market.

Domestic opportunities

Under the background of high industry outlook, opportunities for domestic manufacturers will include capacity demand support and technical capacity follow-up.

There are already traces to follow in the internal of listed companies. Recently, in the process of communication with investors, the senior management of horotel said that the localization progress of the company is faster and faster. In the early stage, some localization of MLCC was carried out, and the proportion of localization of various raw materials will be continuously increased in the future. Since localization requires customer certification, the company will speed up the process after obtaining the certification.

The company also said that in recent years, the market of controller industry has been shifting to China, and the growth rate of orders received in China during the epidemic period has accelerated, but it is not only the impact of the epidemic.

"In fact, in the first half of this year, domestic semiconductor enterprises encountered a once-in-a-lifetime shortage opportunity. Many companies have been given more growth opportunities and some opportunities in overseas customers." For the Chinese market, Gartner expects the revenue share of Chinese semiconductor companies to double from 15% in 2020 to 30% in 2025, Sheng said.

The current proportion is caused by many factors. For example, some OEM products produced in China will not use domestic chips; At the same time, the proportion of domestic chips used by domestic electronic enterprises will continue to increase.

From the perspective of semiconductor industry chain, Sheng Linghai pointed out that there is still a lack of IDM giant companies in China, compared with domestic packaging and testing companies accounting for more.

In addition to the production of OEM links, with the growing domestic semiconductor ecology, attention and expectations for emerging industries are also rising. For example, risc-v, which is regarded by some as a promising "one-third of the world" in terms of architecture, and the third-generation semiconductor market with a small gap between China and overseas.

Sheng Linghai analyzed to the 21st century economic reporter: "frankly speaking, risc-v is still in a relatively early stage. The reason is that risc-v is still a single digit in terms of scale and authorization. Here "single digit" refers to the products made by a company based on this architecture. There may be dozens of companies and hundreds of potential customers, but compared with arm (its size) is still in the early stage. "

He further explained that the most difficult thing to develop is actually the software and hardware ecology. Leading enterprises need to do chip R & D and design around risc-v architecture, so that ecology can be formed. Some people in the industry even think that it will take at least five or even ten years before risc-v is more and more widely used.

Sheng Linghai also pointed out that although risc-v architecture is open source, it still needs to provide mature IP before it can be successfully implemented. If only open source, most companies can't do it, because redesigning an architecture is a resource consuming thing.

From this perspective, it might be a good thing if Intel completes its acquisition of sifive“ At that time, Intel is likely to invest more resources in this area to promote the development of industrial ecology. In the past, arm ecology also needed the support of large companies. With Nokia, Samsung and apple using arm (only gradually improved its ecology), otherwise it would be very difficult. "

As for the third generation semiconductor, Sheng Linghai believes that it is mainly used in power devices and power amplification. There is demand for new energy vehicles, 5g construction and other scenarios, but the number of them is small at present. For example, Tesla does use silicon carbide based devices in new energy vehicles, but it will not be able to replace all traditional silicon-based products soon.

The core lies in the fact that the market is not huge at present. "The global market size of the third generation semiconductor is expected to be about $10 billion," he added. Frankly speaking, at present, there are good enterprises in the field of silicon carbide in the United States and Europe. Domestic enterprises, rather than invest in a scattered way, should concentrate on a small number of enterprises and concentrate their talents on tackling key problems, so as to successfully produce silicon carbide products. We feel that we still need to gather all our abilities instead of dispersing them

?