Zhejiang University couple at the helm, Changguang Chenxin is preparing for an IPO.

Ask AI · How can the overseas experience of Zhejiang University couple help the company’s technological breakthroughs?

The Hong Kong Stock Exchange is about to welcome a pair of top students.

Industry insiders have learned that Changchun Changguang Chenxin Microelectronics Co., Ltd. (referred to as “Changguang Chenxin”) is about to list on the Hong Kong Stock Exchange, backed by the Zhejiang University alumni couple Wang Xinyang and Zhang Yanxia.

Avoiding the red ocean of consumer-grade markets, Changguang Chenxin mainly develops CMOS image sensors for industrial and scientific fields. The name is quite unfamiliar, but the application is very broad—such as automated inspection in factories, or cell imaging under microscopes.

I still remember during the 2022 financing round, the company’s valuation rose to 10 billion yuan, and it once aimed for a listing on the STAR Market, but ultimately shifted to Hong Kong stocks.

A Hidden IPO in Northeast China

Have you ever wondered: when we record videos with our phones or doctors use probes to observe internal conditions, how are these vivid images displayed on screens?

The answer is image sensors. Their principle is not mysterious; fundamentally, it’s an optoelectronic exchange—using photodiodes, photons turn into electrons, which are then read out and converted into a language that computers can recognize, forming the patterns we see with our naked eyes.

The core is simple, but to achieve higher resolution, sensitivity, and frame rate, the entire conversion structure must be very clever. This brings us to today’s main subject—the CMOS image sensor (abbreviated as “CIS”).

In terms of application scenarios, consumer electronics, automobiles, and security dominate most of the CIS market share. Giants like Sony, Samsung, and OmniVision have long held the market, leaving little room for others.

Changguang Chenxin focuses on a very niche area—industrial imaging and scientific imaging, which together account for less than 3% of the entire CIS market. In these two segments, Changguang Chenxin ranks third globally, with market shares of 15.2% and 16.3%, respectively.

What kind of scale is this? From 2022 to 2024 and the first nine months of 2025, the company achieved revenues of 604 million yuan, 605 million yuan, 673 million yuan, and 565 million yuan, respectively. During the same periods, adjusted net profits were 293 million yuan, 223 million yuan, 249 million yuan, and 207 million yuan.

Most of the revenue comes from the industrial imaging sector, increasing from 49.5% in 2022 to the latest 72%. The scientific imaging sector contributed about 26.3% of revenue in the first three quarters of 2025. Professional imaging, medical imaging, and other businesses are almost negligible.

Such a business structure is undoubtedly very concentrated. Once the consumer market was the sweetest pie in the CIS industry, but signs of peaking are emerging. Giants are increasingly turning their attention to non-consumer markets, launching new battles.

Moreover, due to the narrow market, early on Changguang Chenxin heavily depended on orders from the Chinese Academy of Sciences, which accounted for 27.5% of revenue in 2022, making it the largest customer.

Feeling the pressure, Changguang Chenxin adopted a straightforward approach—expanding from high-end to mainstream markets, extending into factory automation, positioning, and barcode reading fields. While customer base widened, these fields have lower barriers and fiercer competition, leading to a gradual decline in gross margin.

According to the prospectus, from 2022 to 2024, the gross margins were 76.2%, 63.5%, and 59.0%, decreasing by 17 percentage points over three years, and only slightly rebounding to 64.2% in the first nine months of 2025. Challenges have just begun.

Behind the Zhejiang University Couple

Changguang Chenxin is highly technical, founded by a couple of top students with impressive backgrounds.

Both hold bachelor’s degrees from Zhejiang University, majoring in applied electronics. During their master’s studies, husband Wang Xinyang went to the University of Southampton in the UK, while wife Zhang Yanxia studied at Conestoga College in Canada, then both went to Delft University of Technology in the Netherlands for doctoral studies.

During his studies in the Netherlands, Wang Xinyang first encountered CIS technology, collaborating with Philips Group on sensor development under his supervisor. After graduation, he joined Cypress and CMOSIS, working as both shareholder and engineer, experiencing an unforgettable career.

Wang Xinyang recalled that many projects at that time targeted industrial, broadcasting, and aerospace sectors. Understanding the needs of domestic industries, he began to consider returning to China. In 2012, he returned to his hometown Heilongjiang, and Changguang Chenxin was established. Zhang Yanxia joined the company the following year.

Specifically, OptoElectronics, Wang Xinyang, and Lingyun Optoelectronics signed an investment agreement to jointly establish Chenxin Limited. The actual controller behind OptoElectronics is the Changchun Optical and Mechanical Institute, known as the “Cradle of Chinese Optics.”

Unbeknownst to many, Changguang Huaxin and Changguang Satellite are also incubated by the Changchun Optical and Mechanical Institute. Industry insiders comment, “Any major aerospace project generally involves the Changchun Optical and Mechanical Institute.” Their experience in optoelectronic system R&D and industry resources are invaluable for the nascent Changguang Chenxin.

Wang Xinyang once said in an interview: “This is a state-controlled mixed-ownership enterprise, with half of the ‘national team’ bloodline, which makes the very capital-intensive CIS industry less daunting.” Fortunately, the results have been promising: just over two years after its founding, Changguang Chenxin has successfully developed multiple top-tier CIS chips, exciting many industry experts.

The venture capital circle often says “investment over the Great Wall,” but it didn’t hold true for Changguang Chenxin—along the way, institutions like Hillhouse, Guotou Zhaoshang, CPE Yuanfeng, SMIC Juyuan, Atomic Venture, Yibin Chendao, Zhongke Chuangxing, Fangguang Capital, and others gathered, with valuation reaching 10 billion yuan in 2022.

However, the Zhejiang University couple still retains a significant stake: before the IPO, Wang Xinyang and Zhang Yanxia together held about 49.53% of the company, with OptoElectronics and Lingyun holding 25.56% and 10.22%, respectively.

In June 2023, Changguang Chenxin submitted an A-share prospectus, but after two rounds of inquiries, the listing was terminated. Leaving the STAR Market behind, they set their sights on Hong Kong stocks.

A Lesson to Be Learned

From the very beginning, Changguang Chenxin carried a pioneering spirit.

“As CMOS device developers, our innate goal is to completely eliminate CCD (charge-coupled device). The future of digital imaging will definitely belong to CMOS; the only question is when CCD will be completely phased out, like being put into museums along with film.”

That was 2015, and things developed almost as Wang Xinyang predicted—Sony announced it would gradually cease production of CCD sensors by March 2017, with CIS dominating the core of various fields.

It’s worth noting that CCD once symbolized high-quality images, with its elegant method of sequentially transferring and converting charge. No one expected that CMOS, which initially grew from low-end markets, would break through bottlenecks through iterative improvements, marking the end of its glorious era.

CMOS didn’t surpass CCD overnight. It’s like trying to fit a stone into an egg—early chip manufacturing levels made it hard for CMOS’s high integration advantages to fully manifest. Only with finer process nodes, such as smaller pixel sizes, did true evolution begin.

Those familiar with chip research know this was a long march—repeated wafer grinding, etching processes akin to meditation, backed by five or even ten years of technological accumulation.

“Disruptive innovation refers to a product or service that initially takes root in simple applications at the bottom of the market, then gradually moves upward, ultimately disrupting existing competitors,” as Harvard Business School professor Clayton Christensen first defined. Edison, Jobs, Bezos—countless experiments and refinements led to those world-changing moments.

A more recent and typical example is Liang Wenfeng, creator of DeepSeek.

Back then, the mainstream narrative in AI was still “more powerful GPUs, more computing power”—who had more GPUs, who would win. Until last year, when DeepSeek was born, using elegant algorithms to topple the dominance of raw computing power.

Seemingly out of nowhere, but actually built on decades of accumulation—Liang Wenfeng’s work on quantum computing at Fanhui and the “Firefly” AI computer series over ten years, gradually building up computing power and algorithm optimization, enabling a leap from financial tools to AGI.

Each generation has its own efforts; it’s the countless days and nights in labs that make everything we see today possible.

It’s worth recalling that the venture capital world still debates “investing in professors or industry veterans,” “technology or market,” “value or survival,” and that a focus solely on technology has been widely criticized.

But looking at the entire history of civilization, all disruptive innovations are ultimately born from those who dream of dancing with circuit diagrams. True entrepreneurial wisdom may lie in building open-minded, complementary teams, uniting members with different talents and interests.

This journey is arduous.