But the world is not absolute, especially for the third-ranked enterprises in the industry, some barely maintain their operations, and even almost forgotten by the industry; Others are full of ambition, get attention and support from many parties, and have a bright future; Others have become industry stars, and the light has almost overshadowed the first two companies. In short, in the complicated semiconductor industry, the "third" is also a mixed bag.
But overall, the industry leader's revenue scale and market share advantages are very obvious. The gap between the market share and revenue of the second-ranked enterprise and the boss is very obvious, while the relationship between the "third" and the second brother of the industry varies according to the different segments.
In the field of CPU, besides Intel and AMD, Taiwan Province's VIA and Bao Dao also produce x86 processors. I wonder how many people know.
VIA was founded in 1992. After several years' development, it acquired Cyrix (then a division of National Semiconductor Corporation) and Centaur in 1999. Since the acquisition of Cyrix, VIA began to set foot in the field of CPU design with x86 architecture, and introduced a variety of processors. Although its performance is not as good as that of Intel and AMD, its specialty lies in its low power consumption, so it can gain a firm foothold in some special markets. In addition, VIA CPU has a distinctive feature, that is, the hardware integrates the function of data encryption/decryption.
At the same time, VIA also provides PC chipsets for Intel and AMD, but with the development of the market, chipsets provided by third parties gradually withdraw from the market, which also makes VIA's presence in the CPU and related chip market weaker and weaker. Now its CPU market share is less than 1%. I can only watch AMD stand alone against Intel under the leadership of "Ma Su". However, I wonder if AMD's strong counterattack against Intel in the CPU field since 20 16 can bring more confidence and motivation to VIA?
In the field of GPU chips, NVIDIA and AMD have dominated for many years. One focuses on high-performance applications, such as servers and cars, and the other focuses on consumer products, mainly personal computers and game consoles. Since AMD acquired ATI in 2005, there is no third independent GPU chip supplier in the market with a certain market share. This reminds us of the era when GPU was born and was in the growth stage more than 40 years ago. Intel, IBM, and TI compete with each other, so I'll come on stage after you sing.
In the upstream of the semiconductor industry, the role of IP suppliers is becoming more and more important, and the importance of mastering the right to set standards and the core assets of the upstream industry has been deeply rooted in the hearts of the people. In today's GPU IP field, the main manufacturers are only Imagination and Arm, and there is almost no third one.
At present, non-artificial intelligence (AI) is the field that attracts the most attention, has the greatest development potential and has the most far-reaching impact on people's lives and work, whether in the pan-IT field or the semiconductor industry. Whoever can launch a chip suitable for its application in the early stage of AI development will undoubtedly have huge business opportunities, which is exactly what NVIDIA has been doing in recent years.
According to the list of the top ten IC design companies in the world in the first quarter of 2020 released by Tuotuo Industry Research Institute, NVIDIA, ranked third, still performed steadily, with an annual revenue growth rate of 39.6% in the first quarter.
NVIDIA's growth in data centers is quite strong, especially its GPU in high-performance AI computing applications, which stands out in the current chip industry. Because of this, the company's market value once surpassed Intel a few days ago and became the third largest semiconductor company in the world. The top two are TSMC and Samsung.
Such a third place with infinite scenery is rare in the history of semiconductor development, especially for an IC design enterprise, because it is a light asset enterprise and there is no chip manufacturing and packaging factory. In this case, it is rare that its market value can surpass Intel and keep up with heavy asset semiconductor companies such as TSMC and Samsung with high capital and technology density. I remember the last time a similar situation happened in 20 13, when the smartphone market was growing "crazily". At that time, Qualcomm seized the peak of the rapid development of smart phones by virtue of the advanced layout of mobile phone baseband and 4G technology, and its market value once surpassed Intel, which was also boundless.
For wafer OEM and packaging test of heavy assets, companies that start early have great first-Mover advantage, as do TSMC and Sun Moon. It is very difficult for latecomers to catch up. It is almost certain that as long as the leader in the segmentation field does not make directional mistakes, it will be difficult for the second and third brothers behind to approach or even surpass them.
To shorten the gap with the industry leader or the second brother, it is often necessary to take some "unconventional" means.
In the global packaging and testing industry, the products of sun, moonlight and silicon in Taiwan Province Province of China have occupied the first and third positions for a long time, and the second brother is the anchor of the United States. In order to enhance revenue scale and industry influence, Chinese mainland Changdian Technology 20 15 acquired Singapore Xingke Jin Peng, which ranks third in the industry.
After that, there was "indigestion". Although the overall scale has improved, the profit level has dropped significantly. After years of adjustment and adaptation, Changdian Technology and Xingke Jin Peng have gradually realized the "chemical reaction", and their revenues and profits have been continuously improved. Turning losses into profits in 20 19, the operating performance in the first quarter of 2020 reached a record high in the same period of history.
As can be seen from the above figure, Changdian Technology, which ranks third in the industry, is relatively healthy in terms of market share and revenue growth. Four years after the acquisition of Xingke Jin Peng, the company continued to attack the third position in the industry.
As the upstream of the packaging and testing industry, wafer foundry plays an important role in the global semiconductor industry, and can even be regarded as the core of the industry.
TSMC has always been a leader in this field, and the positions of the second and third brothers have changed in recent years. Before 20 17, GlobalFoundries was the second in the industry, and Samsung, which was at a disadvantage, always took TSMC as a catch-up target. For this reason, the group separated the wafer foundry division in 20 17, which is why the revenue calculation method of Samsung's wafer foundry business has changed greatly (the revenue of Samsung's mobile phone processor foundry belongs to the wafer foundry division), which makes the market share of Samsung's wafer foundry increase a lot (at present, it is 18.8%), which is much higher than that of Singer's 7.4%, so the latter is decided.
Singh has been full of ups and downs since his birth. Among the four pure wafer foundries, GlobalFoundries has the shortest history (founded in 2009), and it was born out of the traditional IDM company AMD. After a series of splitting, integration, merger and renaming, today's grid core has been formed.
Although the parent company has invested heavily in the Middle East, Singh's profitability has been unsatisfactory. In 20 18, the company announced that it would abandon the research and development of advanced technology below 12nm (excluding 12nm) and focus on the research and development of characteristic technology.
On the one hand, this strategy is to avoid direct competition with TSMC, so as to reduce risks and improve the efficiency of capital utilization. On the other hand, Singh should vigorously develop the characteristic process SOI. In fact, although SOI is not a new technology, it is really a very good and grounded technology. At present, compared with FD-SOI, RF-SOI has been widely used, especially in the RF front-end of mobile communication terminals represented by mobile phones. The development of FD-SOI is relatively slow, which can be said to be a competitive technology that keeps pace with bulk silicon logic technology. Its biggest feature is less leakage and low power consumption.
Today's wafer foundry industry is in a period of great adjustment. In terms of market share, it is difficult to catch up with the first two. At the same time, UMC and SMIC behind them are also eyeing its position. In addition, compared with the difficulty of catching up with the first two companies, it is much less difficult for UMC and SMIC to catch up with Singh. The market share gap between them is not very big. UMC is 7.3%, ranking fifth.
Recently, SMIC landed in science and technology innovation board, and its market value once exceeded 600 billion yuan, greatly enhancing its financing ability in the A-share market. For wafer foundries with high capital density, it is very important to have sufficient and continuous capital supply, and the science and technology innovation board and "big fund" will give them strong support. Because of this, more and more people believe that SMIC will further increase its revenue and market share.
China Merchants Electronics recently published a research report, expressing optimism about SMIC. It believes that the company's R&D intensity and capital expenditure are higher than the industry average. By 202 1, it is possible to approach or surpass UMC and grid without 10nm and more advanced technology, thus coming to the third position in the industry.
SMIC has mass-produced 14nm chips, and 12nm chips have also been introduced for customer verification. At present, N+ 1 and N+2 generation processes are still being studied. Although the government has not defined the nodes of these two generations of processes, according to industry analysis, N+ 1 is about 8nm process, while n+2 is close to 7. However, these are just guesses, and the accurate information is subject to the official announcement of SMIC.
Of course, the third position of 202 1 industry is our good wish, but it is very difficult to achieve it, especially in the sub-field of wafer foundry, where technical barriers are high and it is difficult to catch up with the former in a short time.
label
It can be seen that in the semiconductor industry, it is not an easy task to sit down and sit in the third position in various segments.
Author/Zhang Jiankeya
# Semiconductor chip industry # #ARM will increase the licensing fees of some customers # # Huawei Hisilicon #