Last week, the news of a merger and acquisition caused widespread concern in the semiconductor industry. The protagonists of the merger are Zhilu Capital and ASE. The former acquired four packaging and testing plants of ASE in Weihai, Kunshan, Suzhou and Shanghai for US$1.46 billion.

There is no doubt that ASE’s core business is in Taiwan, China. The company’s move has greatly reduced its business scale and concentrated its resources on the development of core businesses, especially advanced manufacturing processes and heterogeneous chip packaging businesses. As for the acquirer, Zhilu Capital has obtained mature packaging resources, including plant, equipment, technical personnel, and related patents. These are a good asset for the packaging and testing industry in mainland China, which is in the growth period. Has good investment value. It can be said that in this merger, both parties got what they wanted and it was a win-win transaction.

At the same time, such mergers and acquisitions also reflect the “polarization” of the mid-to-low-end and mid-to-high-end businesses in the semiconductor market. This trend has become more and more prominent in the semiconductor M&A market in the past two years, especially For mergers and acquisitions of about tens of billions of dollars or higher, acquisitions and acquired parties with similar or complementary technologies are mostly leaders in various sub-sectors. These high-quality resources are integrated to form or will form a stronger , A combination with a higher level of technology has further widened the gap with the following companies that are catching up, thus presenting a “polarization” situation.

  

A growing pole

2015 and 2016 were the peak years in the history of semiconductor mergers and acquisitions. After that, 2017 and 2018 were much flatter. However, semiconductor M&A activities have been strengthened in 2019. According to IC Insights statistics, in this year, more than 30 semiconductor acquisition agreements were reached, and the total value increased from US$25.9 billion in 2018 to US$31.7 billion, an increase of 22% year-on-year. The total annual value of the semiconductor acquisition agreement in 2019 reached the third highest ever.

In this year, Infineon acquired Cypress Semiconductor for approximately US$10 billion, and Nvidia acquired Mellanox, a supplier of interconnect and network chips, for US$6.9 billion.

As a leading company in analog chips, MCUs and automotive power semiconductors, Infineon’s industry leadership is recognized. Its acquisition of Cypress is to further consolidate and enhance its original position in the MCU and automotive chip fields. The dominant position. Since NVIDIA acquired Mellanox, its influence in the field of high-performance computing and communications has become more prominent, and the concept of DPU has become popular all over the world since then. Today’s NVIDIA, compared with three years ago, the industry’s dominance has been further strengthened, which can be seen from its market value of nearly 800 billion US dollars.

In the next 2020, driven by 5 major acquisitions and more than 10 small transactions, the total amount of semiconductor mergers and acquisitions in 2020 reached a record high of US$118 billion, surpassing the US$107.7 billion reached in 2015.

In July 2020, ADI announced that it would acquire Maxim Integrated Products for 21 billion U.S. dollars. The acquisition was completed in the summer of 2021. This has further improved ADI’s expertise in automotive systems (especially autonomous vehicles), power management, and dedicated products. IC design market share in analog and mixed-signal ICs. This can be seen as a continuation of ADI’s acquisition of Linear in 2016. The company is constantly strengthening its technical standards and market share in high-performance analog chips, and it has the potential to catch up with Texas Instruments.

In September 2020, NVIDIA announced the acquisition of Arm for 40 billion U.S. dollars. The proposal of this merger has a strong sensational effect, because a top IC design company wants to acquire a top IP manufacturer, which is unprecedented in the history of semiconductor development. With the successful acquisition, NVIDIA’s market “dominance” can be imagined. It will break through the concept of a leading IC design company and may even create a new industry pattern and ecology. It is precisely because of this that, out of fear of the formation of a new monopoly situation, the voices opposed to this merger are very strong. Judging from the current situation, the probability of it being approved by the government of the relevant country is very low.

In October 2020, Intel announced that it would sell its NAND flash memory business and 300mm wafer fab in China to SK Hynix in South Korea for US$9 billion. In fact, for Intel, the NAND flash memory business has already become a “chicken rib” project. At this time, it encountered SK Hynix, which is expanding in multiple business areas. South Korean manufacturers seem to have the genes and desire to make themselves large-scale IDMs. Not only Samsung, but also SK Hynix in recent years have also demonstrated this prominently, not only to strengthen its already strong memory business. , But also to invest heavily in non-memory chips, such as logic semiconductors, CMOS image sensors, etc., to further strengthen its leadership. The acquisition of Intel’s NAND flash memory business in China, on the one hand, has consolidated its position in this field, and on the other hand, it has also put more pressure on China’s local NAND flash memory, which is in the growing period, which is already very strong. The single competitor has become stronger, whether it is technology or scale, local companies will face greater challenges.

Also in October 2020, AMD announced the acquisition of Xilinx for approximately US$35 billion, and the transaction was completed at the beginning of this year. AMD’s acquisition of Gaby is clearly aimed at the high-performance computing and communications market. It has the same goal as Nvidia’s acquisition of Mellanox. The goal is to grab Intel’s market share, and at the same time, it hopes to expand into new application markets that have not been available before. Such a merger of “giants” makes the strong stronger, and will continue to widen the gap with the following companies catching up.

Still in October 2020, Marvell Technology announced that it would acquire Silicon Valley’s high-speed interconnect and mixed-signal IC supplier Inphi for US$10 billion. The acquisition will be completed in the second half of 2021. Marvell’s acquisition is also aimed at the high-performance computing and communications market.

Among the five major mergers and acquisitions that occurred in 2020, three of them are aimed at the high-performance computing and communications market. This also shows from one side that if you want to enhance your competitiveness, widen the gap with your opponents, choose high-tech content. The field is very important, and the mass market represented by mobile phones has been difficult to find room for technology and market expansion, and only the technology and market expansion space for the high-performance computing market is large enough. In the next few years, there will probably be a clear “polarization” situation here.

This “polarization” situation is not only reflected in corporate mergers and acquisitions, similar situations also exist in the development and changes of future semiconductor manufacturing processes.

IC Insights believes that as chip feature size scaling continues to slow down, chip designers have also found it increasingly difficult to justify higher costs. Therefore, the pros and cons between advanced and mature processes have become more clear, and the processes used by different companies have become more targeted. This makes various manufacturing processes have room to show their respective advantages.

Under this development trend, according to IC Insights’ statistics and forecasts, the market share of various semiconductor manufacturing processes is developing in a relatively more balanced direction, as shown in the figure below.

  

As shown in the figure, in 2019, the market share of advanced processes below 10nm is only 4.4%, and by 2024, its proportion will increase to 30%. During this time period, the market share of the 10nm-20nm process will drop from 38.8% to 26.2%; the market share of the 20nm-40nm process will drop from 13.4% to 6.7%; however, from the statistics and forecasts Look, the proportion of mature processes above 40nm has not changed significantly in these years.

It can be seen that the proportion of mature processes has remained basically unchanged, while the proportion of advanced processes has increased substantially in the past few years, from less than 5% in 2019 to one-third of the world in 2024, making these “two poles” in a rivalry. state. This is very similar to the mergers and acquisitions mentioned in the previous article. The market share of mid- and downstream companies will not change much. The leading companies have continuously strengthened their technical reserves and competitiveness through mergers and acquisitions and other measures, so as to expand the market space year by year and expand the market. One-pole proportion. In addition, as mentioned above, the mergers and acquisitions of many leading companies are all aimed at the high-performance computing and communications market, because the technology and market expansion space here is huge. From the perspective of the advanced process development shown in the figure above, it is also in line with this development trend, because the main application market for advanced processes in the future is in high-performance computing and communications.

  

China does a good job of “this pole”

Under these development trends, relatively speaking, in the foreseeable few years in the future, the Chinese mainland market should be the pole with a similar and mature process, because the Chinese semiconductor market is still in the growth stage and is still laying the foundation. First of all, we must base ourselves on doing things well in this pole, and then appropriately develop in the direction of the other pole, but we must not rush for success.

Since 2012, Ziguang Group has successively acquired Spreadtrum and RDA, which opened the prelude to the mergers and acquisitions of the semiconductor industry in mainland China. Since then, companies and capitals represented by Ziguang have successively carried out many mergers and acquisitions at home and abroad, and some have been successful. , And some failures have brought a number of high-quality assets to the local semiconductor market. Among them, the acquisition of the original NXP Power RF and Standard Devices Division in 2015 and 2016 was the most. After that, especially since 2019, due to the restrictions of international trade, the possibility of acquiring international high-quality assets only exists at the theoretical level. This requires the semiconductor industry in mainland China to do well what it should and can do now, step by step, lay a solid foundation, and do a good job.

  

The Links:   6DI30AH-050 LM190E08-TLG3