SMIC and TSMC will attack the offensive and defensive wars who will dominate the chip market?

In October of this year, Zhang Zhongmou, the founder of TSMC, said that Intel was involved in the foundry field. When Intel talked about it, it was to push the feet into the pool to test the water temperature. He said: "I believe Intel will find that water is very It’s cold.” The global foundry’s output value in 2015 was as high as 48.891 billion US dollars, which is the lifeblood of Taiwan’s technology industry and financial industry.

Will Intel and TSMC's showdown win or lose? Not to mention Samsung, the war has been quietly opened many years ago. Let us talk about the love and hate of the giants today.

The world's first and the world's largest foundry company, foundry market share of up to 54%. In 2015, the capital amount was approximately NT$259.03 billion, with a market capitalization of approximately US$153.6 billion (2016/9) and approximately NT$5 trillion. On the other hand, TSMC's capital expenditure in 2016 was as high as 9.5 billion to 10.5 billion US dollars (about NT$305 billion to 338 billion yuan), which has surpassed Intel.

The process takes a steady path from 28 nm to 20 nm to the mature Q2 process in 2015 (which can be mass produced and stabilized in terms of efficiency and yield) up to 16 nm. The advanced process of 10 nanometers is expected to be mass-produced in the first quarter of 2017. It also revealed at the end of September this year that in addition to the 5 nanometer process is currently actively planning, the more advanced 3 nanometer process has also organized a research and development team of 300 to 400 people.

SMIC and TSMC play offensive and defensive wars Who will dominate the chip market?

In the future, physical limitations will make the process and Moore's Law more difficult to implement. TSMC is expected to continue to invest in advanced process R&D, but it also focuses on the dual strategy of mature process specialization to maintain its leading position in foundry.

1. How is UMC outdated?

UMC is second only to TSMC and the second largest wafer foundry in the world. However, in 2015, Grofund became the third child with a market share of 9.6% and a market share of 9.3%. In fact, the foundry industry has only one leader. When the economy is not good, only TSMC has always maintained profit, and the remaining 2, 3 and 4 are a group of melee.

Founded in 1980, UMC was also the first semiconductor company listed in Taiwan and has been a leader in foundry since its inception. What caused UMC and TSMC to call the wafer duo, and now, regardless of stock price, revenue and profit, Taiwan Semiconductor Manufacturing Co., Ltd. is not in the foundry position? This is to tell the story of Zhang Zhongmou, Chairman of TSMC, and Cao Xingcheng, the honorary chairman of UMC.

Zhang Zhongmou went to the United States to study in 1949. He obtained a bachelor's and master's degree in mechanical engineering from the Massachusetts Institute of Technology. Because he failed to apply for a doctoral degree, he had to enter Texas Instruments before he graduated. At the time, Zhang Zhongmou was 27 years old. At that time, Deyi was producing four transistors for IBM. IBM provided design and OEM, which can be said to be the prototype of foundry. Zhang Zhongmou led several engineers and successfully upgraded the yield of Texas Instruments from 2% to 3% to more than 20%, even exceeding IBM's own production line.

Zhang Zhongmou spent 25 years in Deyi, until 1983 decided that there would be no opportunities for promotion. In 1985, at the invitation of the Minister of Economy Sun Yunxi, he returned to Taiwan as the Dean of the Industrial Research Institute. At that time, Zhang Zhongmou was 54 years old. Compared with Zhang Zhongmou's foreign academic qualifications and foreign business experience, Cao Xingcheng entered the ITRI after graduating from the Bachelor of Electrical Engineering and the Graduate School of the National Taiwan University. After the establishment of UMC in 1980, ITRI was transferred to the deputy general manager of UMC and transferred to the general manager the following year.

Let us look again - UMC was founded in 1980, Cao Xingcheng was deputy general manager in 1981, and Zhang Zhongmou was the chairman of UMC in 1985. In 1986, Zhang Zhongmou founded TSMC and was also a member of the Industrial Research Institute, UMC and TSMC. Compared with UMC, which develops its own processor and memory products based on integrated component design (IDM), TSMC specializes in foundry.

This was completely a pioneering work at the time, and no one was optimistic. It is generally believed that it is impossible for IC design companies to hand over the chips to outsiders for production and confidentiality. Moreover, the added value created by foundry is more than the sales of chips. Much lower. However, the capital expenditure for establishing a fab is very expensive. If the design and manufacturing of the chip are separated, IC design companies can concentrate their energy and cost on circuit design and sales, while companies specializing in foundry can simultaneously Several IC design companies provide services that maximize the utilization of their production lines and place capital and operations on expensive fabs. The success of TSMC has also prompted the rise of Fabless.

However, this completely angered Cao Xingcheng. He claimed that Zhang Zhongmou had proposed the foundry to Zhang in the year before Zhang returned to Taiwan, but he did not respond. As a result, Zhang Zhongmou took the government in the next year. Resources, and the Dutch-owned Philips, which is based on its own personal relationship, ventured into a foundry company.

At that time, Cao Xingcheng was selected to hold a press conference on the eve of the signing of the contract between the Industrial Research Institute and Philips, and announced that UMC would expand the new plant to compete with TSMC. Since then, Cao Xingcheng and Zhang Zhongmou have not stopped fighting each other; however, Zhang Zhongmou has also served as the chairman of UMC. Until 1991, Cao Xingcheng succeeded in uniting with other directors to avoid the competition and forced Zhang Zhongmou to resign and climb from the general manager. Go to the chairmanship.

The success of TSMC in the foundry of the foundry has also become a reference for UMC. In 1995, UMC abandoned its own brand and transformed itself into a purely professional wafer foundry. Cao Xingcheng’s idea is more savvy than Zhang Zhongmou. He thinks that if he can establish a wafer foundry with a factory-free IC design company, he will not have the money to cover the expensive fabs, and secondly, he will grasp the stable needs of customers. Can directly undertake the orders of these IC design companies. Therefore, Cao Xingcheng developed the so-called "UMC mode" and established joint ventures with 11 IC design companies in the United States and Canada to establish Liancheng, Lianrui and Lianjia foundry companies.

However, the risk of technology outflows accompanying this move, large IC design plants began to be reluctant to give chip design drawings to UMC foundry, making UMC's customer base a large number of small and medium IC design plants. In 1996, because of the customer's questioning the establishment of the IC design department in the foundry, there will be doubts about the misappropriation of customer design. UMC has also divided its IC design department into a company, including the current MediaTek, Lianhe Technology, Lianyang Semiconductor, Zhiyuan Technology and other companies. Then there is the problem that the equipment is not unified. The factory equipment that is joint venture with different companies will be slightly different. When a factory orders burst, it is difficult to transfer to other factories and waste excess capacity.

In contrast, TSMC uses its own funds to build its own factories. It not only allows international companies to hand over advanced processes to TSMC, but also fears that their trade secrets will be stolen and their production capacity will be fully utilized.

However, the real watershed that really allowed Cao Xingcheng to smash the entire Hongtu hegemony and then catch up with TSMC was also a fire in 1997. The cooperation between UMC and IBM failed in 2000.

As mentioned above, each of the UMC's fabs is an independent company. “Lianrui” is another new gossip plant of UMC at that time. More than two years after the completion of the factory, trial production began in August 1997, and the second month of production rushed to more than 30,000 pieces. In October of that year, General Manager of UMC said in an overwhelming tone: "UMC must kill TSMC within two years!" Two days later, a man-made fire destroyed the Lianrui factory. The fire not only destroyed tens of billions of factories, but also allowed Lian Rui to be able to make a profit of 2 billion yuan for UMC. It also missed the peak period of semiconductor boom, and the loss of orders and customers. It is the worst fire damage in Taiwan. Once, it also hit the insurance industry and lost more than 10 billion yuan, which made Taiwan's technology factories and property insurance industry raise awareness of risk control and prevention. This is not to mention.

Yaskawa SGM7A Type Servo Motor

Yaskawa Low inertia, high speed servo motor SGM7A

Yaskawa SGM7A type servo motor

Wuxi Trenty Machinery & Equipment Co., Ltd. ,