CMOS image sensors are about to explode, and domestic companies are waiting in anticipation

Source: The content comes from "GF Securities", thank you.

The working principle of smartphone cameras is that the scene is captured through the lens group to generate an optical image, which is projected onto the image sensor. The image sensor converts the optical image into an electrical signal, which is then converted into a digital signal through analog-to-digital conversion. The digital signal is processed by the DSP (digital signal processing chip) and then sent to the processor for processing, and finally converted into an image displayed on the screen. In terms of physical structure, it is mainly composed of physical components such as the lens group, focus motor, holder/lens mount, infrared cutoff filter, image sensor, PCB board, etc.:

Protective film: mainly protects the lens from collision and scratches;

Prism group: The lens is equivalent to the eyes of the camera module, which determines the quality of light entering and the

Imaging on optical materials. It can be divided into resin lenses and glass lenses. Resin lenses are the mainstream in current smartphone camera modules.

Autofocus unit (VCM): The main function is to realize the autofocus of the camera module. The position of the lens is adjusted by changing the driving current of the VCM to achieve the focusing function. If this component is not present, the camera module is a fixed focus module.

Infrared cut-off filter: It uses precision optical coating technology to achieve high transmittance in the visible light region (400-630nm) and cut off near infrared (700-1100nm). Its main function is to filter out infrared light and ensure that the light reaching the image sensor is visible light, so that the captured image also meets the perception of the eye.

Image sensor (CIS): The core component of the camera module. After light enters the camera module through the lens, an image is formed on the CIS. CIS converts the optical signal into an electrical signal. Almost all current smartphones use CIS based on CMOS technology.

Flexible circuit board: plays the role of line connection and signal transmission in the camera module.

According to Yole's statistics, in 2016, CMOS image sensors accounted for the largest share of mobile phone camera modules by value, accounting for 42.3% of the market value; according to TrendForce's statistics, in 2016, CMOS image sensors accounted for 52% of the market value in mobile phone camera modules.

Yole statistics on the market size of each link of the camera (2016)

The CMOS image sensor industry chain is mainly composed of upstream wafer foundries, packaging companies and testing companies, midstream chip design companies and downstream module manufacturers and end customers. CIS design manufacturers are at the core of the industry chain, and their product solutions are outsourced to wafer foundries, packaging and testing companies for chip manufacturing, coloring, packaging and testing. Qualified products are shipped to end customers (smartphone manufacturers, security monitoring equipment manufacturers, medical equipment manufacturers, etc.) through logistics centers.

Overview of CMOS Image Sensor Industry Chain

From the perspective of industry model, there are mainly two models: IDM and Fabless:

The IDM (Integrated Device Manufacturer) model refers to a business that covers the entire process of chip design, chip manufacturing, and chip packaging and testing. Major manufacturers include Samsung, ON Semiconductor, SK Hynix, STMicroelectronics, etc.

The fabless model means that the company has no production and processing capabilities and only carries out product design work. The design layout is then handed over to the wafer foundry for processing, and then the chips processed by the foundry are handed over to the packaging and testing manufacturers for packaging and testing. Beijing OmniVision and GalaxyCore belong to this model, and the raw material procurement is mainly wafers.

According to Yole Development data, CMOS image sensors are widely used in smartphones, consumer, computer, automotive, medical, security and industrial applications. Smartphones are the main downstream application, accounting for more than 60% in recent years.

The downstream applications of CMOS image sensors are mainly smartphones

From the perspective of demand-side preferences, optical cameras are an important factor that consumers care about. According to a survey conducted by Sino Consulting, in response to the question of the final reason for purchasing existing mobile phones, the proportion of consumers who chose body design (41.0%), camera pixels (27.8%) and photo effects (21.7%) ranked first in the three dimensions of appearance, software and hardware configuration, and operating experience, indicating that consumers attach great importance to optical cameras. This is also the reason why mobile phone brand manufacturers often use smartphone cameras as an important selling point of their products. Therefore, under the trend of stock market and consumption upgrading, optical cameras will become the most important consumer electronics track.

Appearance and camera are the two most important reasons for consumers to buy mobile phones

Looking back at the development history of mobile phone cameras, it has gone through five stages: megapixels from 2000 to 2005, 10 million pixels + initial innovation from 2006 to 2009, high pixel competition from 2010 to 2013, large pixel competition from 2013 to 2016, and dual-camera from 2016 to 2017. It can be seen that the development and upgrade of optics revolves around pixel upgrades and innovative functions. Combining the development history of cameras, technological evolution and consumer preferences, we believe that the upgrade trend of consumer electronic optical cameras will follow the following two paths: one is the technical upgrade at the two-dimensional level, which mainly includes two parts: technical upgrades to make the photo effect close to SLR, and the camera module is small or even hidden to create a full-screen mobile phone; the other is the technical leap from 2D to 3D, realizing the transformation from obtaining two-dimensional images to obtaining three-dimensional information.

Technology tree of optical development

Among them, the upgrading of technology to make the photo effect close to that of SLR is a very important upgrading trend. The photo function of mobile phones has made a huge leap in recent years, but there is still a certain gap between it and professional SLR cameras, which is reflected in the three aspects of image quality (photos taken with SLR are better than mobile phones in image quality, tolerance, color resolution and detail processing), depth of field control (the foreground and background of photos taken with SLR are blurred very naturally) and zoom function (SLR uses optical zoom instead of digital zoom, and the resolution remains unchanged after zooming, and there is no loss in image quality). The reason for the gap between mobile phone cameras and professional SLR cameras is mainly in hardware (such as the size of the photosensitive element of SLR cameras is much larger than that of mobile phone cameras, the aperture configuration of professional SLR cameras is ahead of mobile phones, and optical zoom is used).

However, we believe that although the size and internal space of mobile phones have led to a gap between the hardware and the effects of SLR cameras, the goal of smartphone cameras to approach SLR cameras in terms of hardware and effects is what mobile phone brands are striving for, what consumers are demanding, and therefore the direction of future upgrades. Therefore, the technical upgrades related to SLR-like photography deserve our special attention.

The combination of dual cameras can break through the bottleneck of a single camera and approach the performance of a single-lens reflex camera by using the combination of hardware and algorithms. 2016 also became the first year of the dual camera explosion, and dual cameras also derived different hardware and algorithm configuration solutions. In terms of specific technical solutions, dual cameras can be divided into the following four combinations: parallel dual cameras with the same pixel for high light intake (color + color), stereo cameras with different pixels for scene effects (imaging + depth of field), color + black and white solutions for low light effects, and wide-angle + telephoto solutions for smooth zoom.

In summary, compared to a single camera, the dual camera not only expands the image sensor area, achieves pixel improvement and increases the photosensitive area, but also realizes new functions such as depth of field shooting, optical zoom, fast high dynamic HDR, etc., bringing consumers a better photo-taking experience.

Dual cameras on smartphones

The dual-camera solution is not perfect. Although there are many solutions, different dual-camera solutions achieve different effects. Previously, most dual-camera solutions were in two modes: "wide-angle + telephoto" or "color + black and white". "Wide-angle + telephoto" focuses on optical focus, and achieves SLR-like optical zoom function through algorithms and conversion of two different focal lengths, but has the disadvantage of poor night scene shooting. In the "color + black and white" solution, the color lens is responsible for recording the overall picture, and the black and white lens records the details of the picture due to its high light intake and high pixel characteristics, focusing on image details and imaging in dark light environments, but cannot achieve optical zoom.

The triple-camera solution adds a new camera, combines the advantages of dual-camera, makes up for the shortcomings of dual-camera, and makes the photography closer to that of SLR. The triple-camera solution complements the hardware advantages by configuring three different optical imaging components, and uses software algorithms to achieve functions that dual-camera cannot achieve.

Specifically, from the current technical solutions on the market, the three-camera solution can be divided into three categories, realizing the organic combination of low-light, depth of field, and zoom functions. At present, Huawei's three-camera solution is the earliest to be applied and has been deployed in many mobile phones. Its three-camera solution can basically be divided into three categories: "main camera + ultra-wide angle + depth of field" ultra-wide angle depth of field solution, "color + black and white + telephoto" zoom camera solution to improve low-light shooting and "color + black and white + ultra-wide angle" wide-angle camera solution to improve low-light shooting and "ultra-wide angle + wide angle + telephoto" FishEye zoom camera solution with full functions.

The characteristic of the three-camera solution is the fusion of advantages

CMOS image sensor (CIS) is an important component of optical cameras. Its function is to convert the received optical information into electrical signals, and then convert the electrical signals into digital signals through analog-to-digital conversion, so as to be processed by the mobile phone to output the final image. Its main downstream application areas are smart phones, security and automobiles, etc., and the main upstream raw material is silicon wafers (for the basic knowledge of CMOS image sensors, please refer to the appendix).

Looking ahead, we believe that the demand market for CMOS image sensors will see rapid growth, thereby bringing about the growth of corresponding upstream silicon wafers. On the one hand, we expect the penetration of triple cameras to exceed that of dual cameras, which will directly lead to the demand for CIS and wafers. On the other hand, the increase in the proportion of high pixels is an inevitable trend, and the average size of CIS will also increase. Therefore, the number of CIS chips cut from the same area of ​​silicon wafers will decrease, and more wafers will be needed to produce CIS.

From the supply side, leading manufacturers Sony and Samsung are actively expanding CIS production capacity, and the supply and demand relationship will be balanced in the long run. However, in the short term, the supply and demand relationship, especially for 8-inch products (below 12 million pixels), is still relatively tight.

Triple-camera and other multi-camera models will directly drive the increase in the use of CMOS image sensors. From the perspective of the structure of the optical camera, one camera needs to be equipped with one CMOS image sensor chip. Triple-camera models are directly equipped with three cameras on smartphones, and the use of CIS is three times that of previous single-camera models and 1.5 times that of dual-camera models, which directly drives the rapid growth of CIS usage.

Driven by the trend of multi-cameras in smartphones, the market has also given a high growth expectation for the market size of CMOS image sensors. According to Yole's statistics, the market size of CMOS image sensors reached US$13.9 billion in 2017, a year-on-year increase of 17%, and it is expected that the overall CIS market size will grow at a compound growth rate of 9.4% from 2017 to 2022.

Yole forecasts the market size of CMOS image sensors

Another consulting firm, IC Insights, also predicts that from 2017 to 2022, the year-on-year growth rates of CMOS image sensor sales and sales volume will reach 8.8% and 11.7% respectively. As the first driving force of the CIS market, the growth rate of smartphones will exceed the growth rate of overall CMOS image sensors. According to Yole's statistics, in 2017, the proportion of CIS on mobile phones accounted for about 67.8%, with a scale of US$9.44 billion, and its growth has a significant pulling effect on the overall CIS market.

IC Insights forecasts CIS revenue (left axis) and unit sales (right axis)

We have made a simple calculation here, conservatively assuming zero growth in smartphone shipments. At the same time, based on TrendForce's forecast of the penetration rate of dual/triple-camera models in 2018-2020, we have calculated that without considering other factors, driven solely by the multi-camera trend, the demand for cameras, or CMOS image sensors, will grow by nearly 20% in the next two years, exceeding the growth rate of overall CMOS image sensors.

The rapid growth of downstream demand will also lead to an increase in the use of upstream wafers for manufacturing CMOS image sensors. We expect the growth rate to be consistent with downstream demand, which is about 20%.

In addition to the increase in usage directly driven by the triple camera, we believe that the increase in usage of CIS upstream silicon wafers also comes from the increase in wafer usage due to the increase in the size of a single CIS as the pixel increases:

First of all, the increase in the proportion of high pixels is still an important trend in the future, especially the main cameras of dual/triple cameras are gradually migrating towards high pixels. Currently, Sony and Samsung have also launched 48-megapixel products. Generally speaking, the higher the pixel, the larger the average size of the CIS will be. The result of the increase in the average size of CIS is that the number of chips cut from each silicon wafer decreases, so more wafers are needed to produce CIS. However, it should be noted that the second and third cameras in the dual/triple camera solution do not necessarily have to use high-pixel cameras, so the structural change of pixels is a relatively slow process.

Therefore, the overall demand for CIS usage due to the increase in pixel size will not be particularly high, but it can still bring a certain upward momentum to the growth rate center on the demand side.

Since Huawei launched the triple-camera model P20 Pro in March 2018, equipped with the 40-megapixel Sony sensor IMX 600, smartphone cameras have officially entered the 40-megapixel field. Currently, the top two CIS manufacturers, Sony and Samsung, have released 48-megapixel products, namely Sony's IMX586 and Samsung's GM1:

Sony IMX586: In July 2018, Sony released its first 48-megapixel CMOS image sensor product, which is 1/2 inch in size (8.0mm in diagonal length) and the size of the unit pixel is reduced to 0.8 μm. This sensor is equipped with Quad Bayer arrangement technology (the 40-megapixel Sony IMX 600 equipped with Huawei P20 Pro also uses this technology), which uses 4x4 RGB array imaging and supports the operation of 4 adjacent pixels (traditional Bayer previously only supported 2x2).

Schematic diagram of Quad Bayer arrangement structure transformation

In bright environments such as outdoor during the day, 48-megapixel images can be output, which is clearer than traditional 12-megapixel products. In dark environments (0.8 μm unit pixels are too small to capture enough light), the sensitivity can be increased to 1.6 μm pixel size by adding 4 adjacent pixels (the pixel is 12 million pixels at this time, and the output result is clearer). Currently, smartphones equipped with Sony IMX 586 include: Huawei nova 4, Honor V20.

Samsung GM1: Samsung subsequently released a 48-megapixel CMOS image sensor product in October 2018, which is also 1/2 inch in size and 0.8 μm in unit pixel size. Samsung also expanded the array to 4x4, but the difference from IMX586 is that each 2x2 array can only recognize the same color and can only output data together, so it can also be considered equivalent to a 12-megapixel CIS product with a unit pixel size of 1.6 μm. However, Samsung GM1 can still achieve 48-megapixel photo effects through other methods such as interpolation. The smartphones currently equipped with it are: Redmi Note 7.

Samsung GM1 output diagram

As these two CIS products from Sony and Samsung are already installed in smartphones, the 40-megapixel high-pixel market will continue to penetrate in the future, and the main cameras of dual/triple cameras will continue to migrate towards high pixels.

The upstream of CIS is wafer manufacturing, which is a silicon wafer that cuts out CIS chips of corresponding sizes. Therefore, the increase in the average size of CIS results in a decrease in the number of chips cut out of each silicon wafer, so more wafers are needed to produce CIS. We made a simple calculation in the table below. When the average size of CIS increases from 12 mm² to 20 mm², the wafer usage becomes 1.69 times the original. When it increases from 20 mm² to 25 mm², the usage becomes 1.26 times. When it increases from 25 mm² to 35 mm², the usage becomes 1.42 times.

(Note: Since CIS with 12 million pixels and above is mainly produced on 300 mm production lines, and CIS with less than 12 million pixels is mainly produced on 200 mm production lines, the calculation of the multiple relationship is converted into 200 mm wafer production lines or 300 mm wafer production lines to ensure comparability.)

Calculation of the number of wafers consumed by CMOS image sensors with different pixels

Therefore, as the proportion of high pixels in cameras continues to increase and the average size of CIS increases, there will be more demand for upstream wafers. However, since the increase in pixel proportion is a relatively slow process, we expect that the increase in wafer demand due to the increase in the average size of CIS will not be very large, but there will still be demand, and we estimate that it can increase the growth rate of CIS wafer demand by 2-4 percentage points.

According to IHS data, in 2017, when dual cameras were rapidly penetrating, the overall pixel structure did not change very fast. In 2019, the first year of triple cameras, since the second and third cameras other than the main camera do not necessarily need to use high pixels (for example, the main functions of the telephoto lens and depth of field lens in the current three-camera mobile phones on the market are still zoom and ranging, and the main imaging is still the main camera, so the telephoto and depth of field lenses mostly use 5 million or 8 million pixels), low-pixel products of 800 pixels and below will still appear on triple cameras, and the speed at which the living space of low-pixel products is squeezed will actually be relatively slow, so the speed of pixel structure changes (lower proportion of low pixels, higher proportion of high pixels) will also be relatively slow.

In addition to smartphones, the security and automotive markets are also important application areas for CMOS image sensors. Looking ahead, security and automotive CIS will also see good growth:

Security CIS: The global and domestic security market has huge capacity and will continue to maintain long-term stable growth in the future. As an important device at the front end of video surveillance, cameras are expected to grow in number in the future and develop in the direction of high-end, while boosting the market size of the corresponding CIS.

Automotive CIS: Intelligence is an inevitable trend, and driverless driving will become the ultimate goal of automotive driving. ADAS, as an important basic product in the transition stage, will see a rapid increase in penetration. As one of the key sensors in the ADAS perception layer, the market space for vehicle-mounted cameras will increase rapidly, directly driving the growth of the CIS market size.

The global and domestic security market has a huge capacity and will continue to maintain long-term stable growth in the future. After more than half a century of evolution, the global security market has developed into a mature industry with a large market size. The application areas have expanded from the earliest political and military sensitive fields to commercial fields such as office buildings, hospitals, and schools, and then to the residential field, and the space is constantly expanding. According to statistics from the Prospective Industry Research Institute and China Security Network, the global security market reached US$256 billion in 2017, and the total output value of China's security industry reached 620 billion yuan. In the future, as governments around the world continue to pay attention to security issues, IT communications, biometrics and other related technologies continue to advance, and the upgrading and replacement needs from developed regions in Europe and the United States and the new demand from emerging markets will drive the security market to continue to grow. It is estimated that by 2022, the global security industry market size will reach US$352.6 billion, with a compound growth rate of 6.5%.

In the security industry, security products account for 35% of the market share, while video surveillance accounts for about 50% of the market share. Optical cameras are at the front end of video surveillance, responsible for the collection of audio and video information, and are important basic equipment in the security industry chain. The multi-dimensional, all-weather, three-dimensional and intelligent front-end (perception) is an important foundation for system performance. Security CIS has also maintained rapid growth in recent years. According to Yole's statistics, the market size in 2017 was US$786 million, a year-on-year increase of 26%. Looking ahead, as the security market scale continues to expand, security CIS will usher in growth in terms of quantity.

TSR predicts that the global market sales of security video surveillance lenses will reach 184 million units by 2020, with a future compound growth rate of approximately 4.6%. On the other hand, the high-definition, networked, and intelligent development trend of security video surveillance products will also put forward higher requirements for image quality. The proportion of CIS sensors with high photosensitive area and high pixel count will further increase, which will also further boost the scale of the security CIS market.

Security CIS scale (left axis) and growth rate (right axis)

With the further development of communication networks and the advancement of technologies such as artificial intelligence, driverless cars will have the advantages of improving traffic efficiency and driving safety in the future. Automobile intelligence will be one of the important trends in the future of automobile electronics. Developed countries represented by the United States and Germany have always focused on supporting the development of autonomous driving at the policy level. Japan, South Korea, China, the United Kingdom and other countries have also actively followed up. At the same time, automobile manufacturers are also vigorously promoting driverless cars. Many car companies in the United States, Japan, Europe and China have set 2020 as the year for the practical application of autonomous driving.

ADAS is the foundation of unmanned driving and the pioneer of automobile intelligence. ADAS (Advanced Driving Assistant System) uses sensors installed on the car to sense the surrounding environment, perform system calculations and analysis, and effectively increase the comfort and safety of car driving. It is an important stage in the transition from human driving to autonomous driving. According to NHTSA, unmanned driving can be divided into five stages. In the L0~L2 stage, it is mainly the application and popularization stage of ADAS. ADAS can realize a variety of active safety functions. With the improvement of ADAS penetration and integration, the intelligence level of the car has been significantly improved and transitioned to the L3 level. When unmanned driving technology enters the L3 stage, unmanned driving can be realized conditionally. With the help of mature vehicle networking (V2X), complete unmanned driving will eventually be realized, that is, the L4 stage. Therefore, the popularization and integration of ADAS can not only promote the intelligence of single vehicles, but also is the basic condition for the realization of unmanned driving.

Benefiting from the continuous penetration of ADAS, the space for in-vehicle cameras is vast, and the compound growth rate in the future is high, which will drive the rapid growth of the automotive CIS market. At least 6 cameras are required to install a full set of ADAS functions for unmanned driving. In the future, as the penetration rate of ADAS increases, in-vehicle cameras will extend from high-end models to mid- and low-end models.

According to Yole's statistics, in 2017, with the trend of automobile intelligence, the growth rate of the automotive CIS market size rebounded. In 2017, the automotive CIS market size was US$658 million, a year-on-year increase of 23%. Looking ahead, the agency predicts that the shipment volume of automotive cameras will increase from nearly 50 million in 2017 to more than 83 million in 2020, with a compound growth rate of 20% from 2014 to 2020. The market size of automotive CIS will also usher in rapid growth.

Automotive CIS scale (left axis) and growth rate (right axis)

CMOS image sensors are a technology and capital intensive industry with the following entry barriers:

Technology and talent barriers: The design of CMOS image sensors covers many sub-fields of integrated circuits. The products are complex and require strong professionalism. At the same time, consumers' requirements for resolution, anti-backlight performance, recognition in low-light environments, stability and reliability are also constantly increasing. CMOS technology is becoming more and more complex. Chip design companies need to have comprehensive technical reserves, rapid design capabilities and sufficient technical talents to cope with increasingly complex challenges.

Scale and capital barriers: For CIS design companies in the fabless model, they need to reach a certain scale to be able to carry out in-depth cooperation with upstream major wafer fabs and packaging and testing plants to establish industrial integration advantages. At the same time, Fab companies also need to invest a lot of money and manpower costs in technology and product development in the early stage. For IDM companies, the investment requirements for plant, equipment, manpower, etc. required for wafer manufacturing and packaging and testing are higher, and funds are also needed to maintain effective operations at ordinary times. Larger companies can take advantage of economies of scale.

Customer certification barriers: As the "heart" of electronic products, the stability and reliability of chips will directly affect the quality of downstream products and user experience. Therefore, downstream customers will adopt strict certification measures for upstream chip suppliers. At the same time, the customer concentration in downstream fields such as smartphones is also relatively high. Therefore, large customer resources and certification have also become important barriers to the CIS industry.

Under such high barriers, the industry also presents a highly concentrated competition pattern, with Sony, Samsung and OmniVision being the top three in the industry. According to Yole's statistics, in 2017, Sony, Samsung and OmniVision had market shares of 41%, 19% and 10% respectively in the field of CMOS image sensors, and the three companies together accounted for 70% of the market share.

Competition landscape of CMOS image sensor market in 2017

From the perspective of production, the industry concentration is also high, with Sony, Samsung and TSMC accounting for more than 70% of the market share. According to Yole's statistics, the number of CMOS image sensor wafers (equivalent to 12 inches) produced in 2017 reached 2.422 million pieces, a year-on-year increase of 2.3%. Among them, the market share of the number of wafers produced by Sony, Samsung and TSMC was 38%, 20% and 16% respectively, and the three companies accounted for 74% of the production share.

CIS chip production (left axis) and growth rate (right axis) of each manufacturer from 2015 to 2017

Currently, Samsung and Sony have a clear attitude to increase CIS production. It is expected that the expansion of leading enterprises will make the compound growth rate of CIS output reach 18%-19% from 2017 to 2020:

Samsung aggressively expands CIS production: Digitimes Research pointed out that Samsung's CMOS image sensor production capacity at the end of 2017 was 45,000 pieces/month. According to Korean media etnews, Samsung's DRAM production line No. 11 in Hwasung, South Korea, was converted to an image sensor production line at the end of 2017 and is expected to be completed by the end of 2018. After the conversion of line No. 11, line No. 13 in Hwasung will also be converted from a DRAM production line to a production line for image sensors. Samsung's total future production capacity will reach 120,000 pieces/month. We expect to achieve the production target in 2020, and Samsung's CIS production will grow at a compound annual growth rate of about 40% from 2017 to 2020.

Sony actively follows up on capacity expansion: According to Digitimes Research, Sony's monthly production capacity in 2017 was about 85,000 pieces. According to Korean media etnews, Sony has increased its CIS production capacity to 100,000 pieces/month in March 2018. In addition, Digitimes Research also shows that Sony hopes to further expand its CIS production capacity to 120,000 pieces/month in 2020. Based on this, it is estimated that Sony's CIS production will have a compound growth rate of about 14% from 2017 to 2020.

Some other manufacturers have a certain willingness to expand production. For example, SK Hynix has been increasing its investment in the CIS industry since 2016. However, due to the small share of production, the impact on the growth rate of production capacity of the entire supply market is also small. We conservatively assume that the compound growth rate of production of other manufacturers is 0, and the production capacity of Sony and Samsung in 2020 will reach 120,000 pieces/month (equivalent to 12-inch wafers). After calculation, the compound growth rate of the entire market from 2017 to 2020 is 17%. After considering the willingness of other manufacturers to expand production, we expect the compound growth rate of the supply side to be between 18% and 19%.

Domestic manufacturers are also stepping up their layout.

First, let’s look at OmniVision Technologies;

Beijing OmniVision Technology Co., Ltd. (hereinafter referred to as "Beijing OmniVision") was formerly known as OmniVision Technologies, Inc., a well-known American semiconductor company founded in 1995. OmniVision is a leading digital imaging solution provider that mainly designs and sells high-performance semiconductor image sensors. It is one of the world's three leading image sensor suppliers along with Sony of Japan and Samsung of South Korea. OmniVision Technologies' global mobile phone, automotive, and security CIS market shares are ranked third, second, and first in the world, respectively.

In May 2015, OmniVision was acquired by a consortium consisting of CITIC Capital, Beijing Qingxin Huachuang and Jinshi Investment for US$1.9 billion. It was finally privatized in early 2016 and became a wholly-owned subsidiary of Beijing OmniVision. At that time, Beijing OmniVision's shareholders were Kaiyuan Suzaku (Shenzhen) Equity Investment Partnership, Seagull Holdings Hong Kong Limited, Seagull Holdings Cayman Limited, Shenzhen Aosi Jiachuang Equity Investment Partnership, and Beijing Integrated Circuit Design and Packaging and Testing Equity Investment Center. After the privatization was completed, Beijing OmniVision transferred its equity several times. At present, its shareholder structure is relatively scattered. The top shareholders are Jiaxing OmniVision, Qingdao Rongtong, Seagull Strategic Investment A3, Xinneng Investment, Jiaxing Shuimu, Jiaxing OmniVision, Shanghai Tangxin, etc. The other shareholders hold less than 5% of the company's equity.

In December 2018, Will Semiconductor released a major asset restructuring plan, intending to purchase 85.53% of Beijing OmniVision's equity by issuing shares. After the acquisition is completed, Will Semiconductor will hold 89.45% of Beijing OmniVision's equity, and Beijing OmniVision will become a subsidiary of Will Semiconductor.

Beijing OmniVision has a multi-industry layout, a leading position, and excellent market share in various fields. OmniVision CMOS image sensor chips are widely used in consumer and industrial applications, including smartphones, notebooks, webcams, security monitoring, automobiles and medical imaging systems. The company's mobile phone CIS market share ranks third, second only to Sony and Samsung. Benefiting from the multi-camera trend, its main applications are auxiliary functional lenses. The automotive CIS market share ranks second, second only to ON Semiconductor. The company's future growth will continue to benefit from the explosive demand for CIS sensors for mobile phone multi-cameras and automotive ADAS system upgrades. The security CIS sensor market share ranks first in the world, accounting for 56%.

OmniVision's market share in various sub-sectors of CCD/CMOS image sensors

Next, look at Sipix Microelectronics;

Founded in 2004, Superpix Microelectronics specializes in the design and sale of CMOS image sensors and image processing chips. The CMOS image sensor chips developed by the company are used in smart phones, tablets, wearable devices, security monitoring, smart cars, robot vision, medical imaging, somatosensory interactive games and other mobile Internet, Internet of Things, special equipment and other fields. Major customers include Lanboco Enterprise Development (Hong Kong) Co., Ltd., China Electronics International Co., Ltd., Shenzhen Hongsheng Investment Development Co., Ltd. and other sales agents. According to the announcement of Will Semiconductor, the company achieved operating income of 460 million in 2017, most of which came from CMOS products on mobile phones.

Sipix's revenue (left axis), net profit (left axis) and gross profit margin (right axis)

From the perspective of industry status, SuperPix has certain advantages in the mid-to-low-end market. Based on its own core technology, SuperPix has successfully developed a number of domestically leading high-performance image sensor chips. Before 2014, SuperPix's CMOS image sensor chips mainly involved the mid-to-low-end field, including 80,000, 300,000, 1.3 million and 2 million medium-pixel CMOS chips. Since 2015, SuperPix's 5 million pixel and 8 million pixel mid-to-high-end products have been launched on the market, and the market scale has gradually expanded.

In December 2018, Will Semiconductor Co., Ltd. released a major asset reorganization plan, intending to purchase 42.27% of the shares of Sipix and 79.93% of the shares of VisionSource by issuing shares (VisionSource is a holding company whose main asset is its 53.85% stake in Sipix. Will Semiconductor Co., Ltd. intends to indirectly acquire 43.04% of the shares of Sipix by purchasing 79.93% of the shares of VisionSource). After the acquisition is completed, Will Semiconductor Co., Ltd. will directly and indirectly hold 85.31% of the shares of Sipix.

Let’s look at GalaxyCore again;

Founded in 2003, GalaxyCore Microelectronics is mainly engaged in the design, development and sales of CMOS image sensors, LCD Drivers, high-end embedded multimedia SOC chips and application systems.

GalaxyCore CIS started from PC cameras and entered the mobile phone field in 2007. With the help of the rapid growth of Chinese mobile phones, it quickly occupied the market. Its products cover from 2 million pixels to 13 million pixels.

According to the official website of GalaxyCore, the company successfully developed the first 2-megapixel and 5-megapixel CMOS image sensors using back-illuminated technology in 2013 and put them on the market. In 2015, the company successfully developed the first 8-megapixel and 13-megapixel CMOS image sensors and put them on the market.

In addition, a series of domestic manufacturers such as SmartSens Electronics and BYD Microelectronics are deeply cultivating this market, waiting for a new round of outbreak.

*The content of the article represents the author’s personal opinion and does not represent Semiconductor Industry Observer’s agreement or support for that opinion.

Today is the 1892nd issue of content shared by "Semiconductor Industry Observer" for you, welcome to follow.