Digital Eyes: China Bids to Capture the Miniature Camera Market

About the Series

China’s recent tech pivot marks a dramatic departure from the stellar run of the private sector-driven platform economy over the past long decade. In this new era, hardware reigns supreme. Toward that end, Beijing is already realigning investments, talent, and institutions, manifest in the very visible hand of industrial policy.

While much attention has been devoted to China chasing the technologies of tomorrow—from advanced chips to quantum computing—what’s equally important to Beijing is securing the technologies of today.

The latter is the focus of our “Tech Pivot in Practice” series because tech supply chains aren’t built with frontier technologies but rather with relatively mature, non-leading-edge technologies that have immediate commercial applications.

These technologies are crucial inputs into end products that few pay attention to because they don’t make the front pages. Yet these are the arenas in which China can dethrone market incumbents without having to invent the future. It simply has to deploy the present at scale to outcompete rivals on cost and market share.

The pivot’s success is far from assured and ultimately rests on the technologies, companies, and people behind it. This series examines the pivot’s peril and promise through these prisms.

The first of the series looked at material science’s rising importance in the tech pivot as China aims to master the silicon carbide power chip. In this second case, we examine the miniaturization of hardware, particularly cameras, and China’s bid to become the dominant supplier to the automotive industry.

Introduction

The transformative impact of the miniaturization of hardware can be seen everywhere. Computers now sit on your wrist or fit in your pocket, transistors are microscopic, and 3D printers rest on a desk. Even the failed Theranos was aiming to miniaturize large blood testing machines.

What’s remarkable is that miniaturization has not come at the expense of capability. NASA’s Apollo 11 computer system, for example, is an abacus compared to the latest iPhone. This inverse relationship between size and functional capacity is equally apparent in the camera.

Beyond just the megapixel race of smartphone cameras, the miniaturization of camera image sensor technology is reaching new heights. That complex and intricate chip has use cases across a diverse range of industries—from consumer electronics and aerospace to automotive and nuclear power.

In this case study, we focus on the current and prospective growth markets for miniature cameras: smartphones and autos, respectively. Although camera sensor technology was once difficult to master, recent progress in image sensor fabrication has enabled China to become a global contender. As Chinese suppliers become increasingly competitive, that will naturally reduce its foreign dependence, particularly on Japanese and South Korean suppliers, and create another significant technology export for China.

Innovation in Image Sensors

An image sensor is a small chip in a camera that converts light waves into signals that are used to capture an image—enabling everything from a Polaroid to an iPhone selfie camera. Two types of sensors—charge-coupled device (CCD) and complementary metal-oxide semiconductor (CMOS)—have jostled for market share.

Both technologies were invented in the United States in the 1960s, but CCD became the industry benchmark for decades because it produced superior image quality and was preferred by high-end professional cameras. The better quality was the result of a CCD sensor chip in which each component is streamlined to work together for optimal output and functionality, from sensing and transferring light into a signal to storing that signal as an image.

In contrast, early CMOS sensor chips produced worse images with more noise because they lacked optimized components and each individual pixel had to do its own complex conversion from light to image. In the decades following its invention, Japanese companies found that CMOS technology was mainly useful for the small integrated chips needed in gadgets like handheld calculators and wrist watches rather than cameras.

Gradual improvements were made, but the breakthrough in CMOS image sensors came in the 1990s when NASA scientists at the Jet Propulsion Laboratory came up with an image sensor initially intended to withstand the harsh conditions of space.

That image sensor was based on a novel concept of CMOS active pixel-sensor (CMOS APS)—a chip that used more advanced lithography to integrate its image sensor and the signal processing circuitry. Led by Dr. Eric Fossum, his team at NASA had stumbled upon a commercial product that could produce images of at least as high quality as CCD and was cheaper to manufacture because it could be made on standard silicon production lines.

Recognizing that potential, Fossum founded the Photobit Corporation in 1995 to commercialize the product, which received huge market validation when American chip giant Micron Technology acquired his company in late 2001. What Fossum may not have realized then was that CMOS would become the key camera technology of the smartphone age. By the end of the last decade, CCD was being written into obsolescence, as the market gave CMOS the vote of confidence as the image sensor technology of the 21st century (see Figure 1).

Figure 1. There’s Only One Image Sensor Now: CMOS
Source: IC Insights; Topsperity Securities Research Institute.

Late Adopter Advantage

Because the mass adoption of CMOS came in the late 2000s, Chinese players have been able to leapfrog into a more mature technology to catch up. China is essentially a giant late adopter of many technologies, which gives it two main advantages.

One, it doesn’t need to invent these technologies, which means it can avoid spending lavishly on basic R&D. Its companies simply need to ride on the prevailing technology trends and deploy them on a grand scale. That can be seen in everything from solar panels and wind turbines to batteries and smartphones. Two, a maturing industry means clear winners are emerging, which allows Chinese investors to better spot and acquire such assets to immediately become a significant player.

China did both when it came to CMOS image sensors, significantly reducing the time needed to catch up.

Case in Point: That Computer in Your Pocket

The case of China’s smartphone industry is about as emblematic of the late adopter dynamic as any. With smartphones came the demand for cameras—in the 15 years since the launch of the iPhone in 2007, no feature has received more attention and upgrades than the camera. Indeed, the “phone” is perhaps the least used feature of the smartphone, which is essentially a pocket computer with a camera that occasionally makes calls.

Not only have smartphone cameras gotten much better in quality, they’ve also grown in quantity. These days, a smartphone comes with an average of four cameras, more than double the number just five years ago. And image sensors are the most expensive component of the camera, accounting for around 52% of the total cost.

It is no surprise, then, that the smartphone segment makes up more than 70% of the nearly $22 billion global image sensor market (see Figure 2).

Figure 2. Smartphones Currently Dominate the Image Sensor Market (2022) Source: Counterpoint.

As image sensor sales took off on the back of the smartphone explosion of the last decade, incumbents like Sony and Samsung made the switch to CMOS sensors and eventually became suppliers to higher end segments like Apple’s iPhone and Samsung’s Galaxy.

These premium devices dominated the first part of the smartphone era but by 2012, Chinese companies entered the market en masse to meet the global demand for affordable smartphones. The results have been staggering: Chinese smartphones grew from 10% to 40% of global market share in just seven years (see Figure 3).

Figure 3. Chinese Smartphones Storm the Global Market in the Mid-2010s (%)Note: Chinese companies are in orange.
Source: IDC.

During these boom years, China’s homegrown smartphone brands needed cameras and mostly imported them. But of course, Chinese suppliers saw a huge opportunity to provide low to mid-range image sensors to companies like Vivo and Oppo. In other words, existing domestic demand for image sensors gave Chinese suppliers a long runway to scale operations and mature their businesses.

And scale they did. Shanghai-based Galaxycore began operations in 2003 but reached new heights in recent years, ranking first globally in terms of units shipped and fourth in revenue. Another Chinese player, Smartsens Technology, was founded as recently as 2017 and has already surpassed Japan’s Canon in sales in just three years.

Why Rent When You Can Own?

Although Chinese companies like SmartSens have risen rapidly, they still trail far behind incumbents like Samsung and Sony. But the obvious standout in China’s column is Omnivision, which was the result of an acquisition by prescient Chinese investors. That acquisition immediately put China into the major leagues of image sensors, as Omnivision ranks among the top three in terms of global market share (see Figure 4).

Figure 4. Chinese Companies Climbing the CMOS LadderSource: Counterpoint.

Table 1. Major Chinese Smartphone Brands Rely on Omnivision
Source: Hongta Securities.

With China’s booming consumer electronics industry, the value proposition of Omnivision seemed obvious (see Table 1). What’s more, just a year earlier in 2014, the Chinese government had established the “Big Fund” to pile money into chips development and issued a National Integrated Circuit Industry Development plan to incentivize private investors to get in on the action.

So by the end of 2015, the Chinese consortium acquired Omnivision, with CFIUS approval, for about $1.9 billion. The consortium didn’t hold on to the asset for long, however, as it sold it to Shanghai-based Will Semiconductor in 2019 for $2.2 billion. That appeared to have been the consortium’s plan all along—to spin the asset off to a company like Will Semiconductor that specializes in chips for automotive applications, the future growth area for miniature cameras.

New Camera Demand: EVs with Eyes

While the smartphone segment will continue to drive camera demand in the foreseeable future, its peak has already come and gone (see Figure 5). With a saturated smartphone market, the mobile device image sensor market is expected to grow at just 3% through 2025. Image sensor suppliers see the writing on the wall and need to find new sources of demand.

Figure 5. Slumping Smartphone Sales Will Affect Image Sensor Demand (in millions)Source: IDC.

That demand is to be found in the automotive segment, especially as electric vehicles (EVs) are evolving into “computers on wheels,” loaded with software for connectivity and semi-autonomous driving. For the computer to drive a car safely and properly, it needs to replicate the two human features that a computer does not have: eyes (cameras) and brain (chips).

So the more autonomous a vehicle is, the more “eyes” are needed, up to dozens of image sensors per vehicle (see Table 2). In fact, the average number of image sensors per vehicle is expected to rise from 2.2 to 7 over the next five years, implying the industry expects most EVs to be able to handle Level 2 autonomous driving. That’s why sales of image sensors for the automotive segment are projected to see 14.3% compound annual growth through 2025.

Table 2. Image Sensor Demand Increases with Automation
Source: Hongta Securities.

Autonomous driving also has policy backing from Beijing. China’s National Intelligent Networked Vehicle Innovation Center released an ambitious plan in November 2020 that aims for 50% of new vehicle sales to be capable of L2 or L3 autonomy by 2025, with a goal of reaching 70% by 2030.

Figure 6. Private Sector Estimates 39% of Vehicle Sales Will Have L2/L3 AutonomySource: Roland Berger; Zhongtai Securities.

Whether such official goals are attainable is debatable, as the discrepancy with private sector estimates attests (see Figure 6). Nonetheless, China’s homegrown EV companies have been charging ahead with making smart vehicles and integrating cameras into their latest models (see Table 3). Similar to the smartphone segment, China also holds a distinct advantage here as the world’s largest EV market, which will create more domestic demand for image sensors as it continues to embrace automation.

Table 3. Chinese Automakers Are Integrating More Cameras into Their EVs
Source: Zuosi Auto Research; Zhongtai Securities; company websites.

These trends bode well for Will Semiconductor, whose acquisition of Omnivision was precisely in anticipation of becoming a competitive supplier to the automotive industry. Omnivision’s ability to offer customizable products that can withstand extreme and variable motion, weather, and light conditions is partly responsible for its success. For example, you don’t want the cameras to malfunction while your EV is on autopilot driving through a Noreaster storm on the dark and curvy backroads of Vermont.

That specialization appears to be bearing fruit as Omnivision’s revenue from automotive increased 85% from 2020 to 2021. It now occupies nearly 30% of global market share in automotive image sensors, trailing only US-based supplier Onsemi (see Figure 7).

Figure 7. Omnivision Is Even More Competitive in Automotive Image SensorsSource: ICV Tank; Zhongtai Securities.

Beyond Omnivision, the strength of China’s domestic EV market suggests that there will be room for homegrown suppliers to develop a supply chain ecosystem. And as China’s EV industry becomes more globally competitive, so too will these image sensor suppliers raise their game and improve along with them.

Different countries will move toward autonomous driving at different speeds, but the direction appears to be the same. So long as the future of cars is one of less human intervention behind the wheel, those cars will need more eyes to instantly capture the 360-degree environment in real time. And it appears that a number of Chinese companies are poised to supply the market with these digital eyes.

AJ Cortese is a senior research associate at MacroPolo. You can find his work on industrial technology, semiconductors, the digital economy, and other topics here.