Competitor Intel is constantly moving, Samsung's technology is gradually advancing, and big customer Qualcomm is not holding back, plus the Chinese government supports local companies; three bad balls and full bases, how can you not lose points? It is strictly testing Zhang Zhongmou, the most senior manager in the world's semiconductor industry.
Last year, TSMC paid the most beautiful report card in history, with a net profit of 263.899 billion yuan after tax, and EPS (pure profit after tax per share) as high as 10.18 yuan; however, starting from the second quarter of this year, TSMC began to face turbulence, from rivals and customers to China, the external situation is not good for this global semiconductor giant. Chairman Zhang Zhongmou faced this "the last battle of the old coach", how to defend and stick to TSMC's wafer foundry industry is probably a difficult but brilliant show.
The first thing TSMC has to face is the challenge of its competitors. In the past, TSMC's opponents were GLOBAL FOUNDRIES, SMIC or UMC, but these companies are no longer threatening; now the real enemies are Intel and Samsung, and these two big manufacturers, such as the "800-pound gorilla", have become increasingly active recently, making Zhang Zhongmou, who has always been calm, inevitably restless.
Challenge 1: Competitors are entering
Intel poaching Samsung process
First of all, Intel invested in Unigroup for US$1.5 billion last year and held 20% of the shares. It not only formed an alliance with China's largest IC design group Spreadtrum and Ruidiko, but will also introduce its products to Intel in the future. In June this year, Intel announced that it would spend US$16.7 billion to buy Altera, which is also one of the important customers of TSMC's advanced process. Intel has been poaching TSMC's wallet, making TSMC's internal enemy.
Intel's strength in the semiconductor industry is undoubtedly. Although its performance in mobile communication chips in the past few years is limited, the CEO Brian Krzanich, who took office last year, has now stepped up its participation in mergers and acquisitions and investments to make up for the weaknesses of mobile chips, and its performance is indeed remarkable.
In addition, Samsung has also shown its efforts in the memory industry in the past to accelerate the technological process. It recently announced that it will officially include 10 nanometer FinFET (fin field-effect transistor) in the development blueprint of Samsung's semiconductor semiconductor process technology. It is scheduled to be mass-produced by the end of 2016, leading the progress, becoming the world's earliest wafer foundry industryman to mass produce 10 nanometer processes.
In fact, Samsung, TSMC and Intel are becoming increasingly fierce in technical competition. Intel recently announced that it will postpone the 10-nanometer period by about half a year and will not be mass-produced in the second half of 17, which will also allow TSMC, which is scheduled to be mass-produced in the first quarter of 17, to win the 10-nanometer competition.
However, TSMC's "leading" voice lasted less than a week, and Samsung immediately announced that it would be ahead of TSMC for at least one quarter, and it was a strong call to TSMC; of course, whether this claim can be achieved still needs to be examined by actual results.
As for the actions of another competitor, SMIC, we cannot underestimate it. Recently, SMIC announced that it will combine Qualcomm, Belgian Microelectronics R&D Center (IMEC) and China's telecom manufacturer Huawei to jointly organize a new technology R&D company to build China's most advanced R&D platform to narrow the technology gap with international competitors such as TSMC, Intel, and Samsung Electronics. The announcement of this joint venture gives SMIC the opportunity to become a first-tier manufacturer.
In fact, SMIC, which has been established for 15 years, has never been able to make breakthroughs in technology. Now, with the support of the Chinese government, on the one hand, it restrains Qualcomm with anti-monopoly and allows Qualcomm to promise to join the SMIC platform; on the other hand, it has brought Huawei, which has the most advanced telecommunications and communication technology. Huawei has a promising HiSilicon Semiconductor, which has a strong potential, will be successfully built in five years, SMIC will have the opportunity to jump to the table and compete with TSMC.
Challenge 2: Large customer structure loose
Qualcomm changes hearts and turns to Samsung, SMIC
TSMC's second biggest challenge comes from customers. The loyal customer structure in the past seems to show signs of loosening, and the most obvious one is Qualcomm, the largest customer.Gaozhuan has always been TSMC's largest customer, but recently encountered MediaTek's continuous market division, and some unpleasant cooperation with TSMC has begun to significantly transfer orders to customers such as Samsung and SMIC. For example, Qualcomm's Snapdragon 820 chips are likely to use Samsung's 14-nanometer process. The big customer Qualcomm Pipa should not be hugged, but the Samsung on the left and the SMIC on the right makes Zhang Dashuai heartbroken the most.
However, Qualcomm has had a lot of trouble recently. Although the second quarter's performance reached its estimated value, the company's estimated revenue in the third quarter fell by 2 to 19% compared with the second quarter, which is significantly lower than the average growth estimate of the foreign investment circle by 4.8%. In addition, it will lay off 4,500 employees, accounting for 15% of the company's employees, and the company may be split. If the situation really evolves like this, it will be a major variable affecting the semiconductor industry.
First of all, according to analysis by many foreign investors, if Qualcomm decides to split its patent license business and chip business, it may sell its chip business to Intel or Samsung, which will reverse the semiconductor industry.
Cowen Co. analyst Timothy Arcuri said that if Intel buys Qualcomm chip business, it will be of great help to the development of Intel's mobile device communication chips and allow Intel to gain a foothold in the Chinese market. In addition, if Qualcomm takes the opportunity to deepen cooperation with Intel and use the latter wafer fab foundry, it will hope to lower Qualcomm chip production costs and improve product price competitiveness, which will cause a painful blow to TSMC, which has lost its Snapdragon 820 chip foundry order, and will also have a great threat to MediaTek, which has originally price advantages.
In addition to Qualcomm, TSMC's client accelerated merger, and the scale became increasingly larger after the merger, which also made TSMC lose its dominant force in negotiating with customers. In March this year, NXP acquired Freescale for US$11.8 billion, and in May this year, Avago acquired Broadcom for US$37 billion, setting a new record for the highest merger and acquisition amount in technology history.
If Intel acquired Altra for US$16.7 billion in June this year, the total amount of global semiconductor companies' mergers and acquisitions this year will exceed US$100 billion. These companies used to be TSMC customers. When the customers became larger and larger, their strength to invest in R&D and even production increased greatly. Whether the orders invested at TSMC in the future is also a key point worth observing.
As for the third challenge, it comes from China. Mainland China has recently shown its determination to use national power to promote semiconductors. From IC design, manufacturing OEM to packaging and testing, there are "national teams" appearing, and they have used large amounts of funds to acquire overseas, which shows that the importance of Chinese customers is increasing, and makes China the most important market in the semiconductor industry in the next ten years.
Challenge 3: China's "national team"
Government support, the market territory will be reorganized
Therefore, investing in China in the past has not been Zhang Zhongmou's option, but now it is obvious that the policy has quietly changed. In December last year, when the Institute of Industrial Research Institute held an academician meeting, Zhang Zhongmou mentioned that he was already thinking about the possibility of landing in and expanding the factory. The reason is that TSMC's orders from China are increasing, not only accounting for more than Japan, but also about the same as the European market. It is expected that in a few years, its importance will exceed that of Europe.
"Even if TSMC does not set up a factory in China, South Korea will go. Samsung has set up a memory factory in Xi'an, Shaanxi." Zhang Zhongmou also implicitly stated, "The Chinese government is playing sticks and carrots, but TSMC has technology that SMIC does not currently have. SMIC's process is N-3 (N represents the most advanced process, N-3 means three generations behind). If TSMC set up a factory, it is expected that the process two generations behind (N-2) is enough."
These three major challenges are coming one after another, allowing Zhang Zhongmou, the ace pitcher who dominated the whole game in the past, to face a situation where three bad balls are full of people. If a bad ball is accidentally shot next, the opponent may go back to home base and score. At the TSMC Act meeting in mid-July, Zhang Zhongmou attended the meeting in particular. Although the briefings were handed over to the co-executive and the Treasurer, he still chose to answer key questions, which shows the Marshal's cautious attitude towards responding to the situation.
However, even in the face of the three bad balls, TSMC has still had a big harvest in recent years, that is, to include Apple in its customer base. When the global smartphone market entered its mature stage, Apple fans remained high loyal to the iPhone, allowing the Apple brand to stand out. Although A8 and A9 chips have adopted risk diversified strategies for TSMC and Samsung, being able to squeeze into the Apple supply chain is a good goal for the company's development. How can Zhang Zhongmou help TSMC invest a few good goals before retirement in the future, and it will also test this veteran coach in the world's semiconductor industry.
