With CNY 2.5 billion in revenue and a monthly capacity of 63,300 wafers, CanSemi Semiconductor has carved out a path in specialty processes within the gaps left by industry giants.
Recently, CanSemi Semiconductor, known as "Guangzhou's First Chip," officially listed on the ChiNext board. With an issue price of CNY 12.01, its shares surged 274.69% on the first day of trading, pushing its market capitalization to CNY 129.5 billion.
At the table of the global wafer foundry industry sit several giants: TSMC, Samsung, SMIC, UMC, GlobalFoundries, and Hua Hong. Their massive monthly capacities of hundreds of thousands of wafers and capital expenditures of tens of billions of dollars have built scale barriers that are nearly insurmountable for outsiders. Yet, within this landscape, gaps always exist: in specialty process fields such as analog chips, power devices, and optoelectronic integration, the market is fragmented and demands are diversified. Large companies cannot cover everything, while small companies cannot handle the volume.
CanSemi Semiconductor is exactly what has grown from this gap. Founded in Guangzhou in 2017, it has achieved CNY 2.5 billion in revenue and a monthly capacity of 63,300 wafers over nine years, and is now officially entering the capital market.
Finding a Path in Specialty Processes Within the Gaps Left by Giants
According to statistics from SEMI and other institutions, the scale of China's wafer foundry market was approximately USD 14.3 billion in 2024, a year-on-year growth of 20.17%. It is projected to reach USD 26.6 billion by 2029, with a compound annual growth rate of 13.22%. Beneath the halo of advanced nodes in the global foundry market lies a larger fundamental base: mature nodes. Markets with considerable growth rates, such as automotive electronics, industrial control, IoT (Internet of Things), and power management, do not require the most advanced nodes; instead, they need specialty processes that are sufficiently deep and specialized.
This is precisely the track chosen by CanSemi Semiconductor. Focusing on analog and mixed-signal chips, it has formed eight process platforms: MS, HV, CIS, BCD, eNVM, MOSFET, IGBT, and SiPho, covering nodes from 180nm to 55nm. In terms of process metrics, this is an entirely different game compared to the advanced nodes of TSMC and Samsung. CanSemi Semiconductor does not hide this in its prospectus, acknowledging that there are still gaps with international giants in process nodes, process accumulation, brand awareness, and operational scale.
However, the foundry business has never been solely about process numbers. The performance of analog chips relies more on process innovation and device optimization rather than dimension shrinkage. What customers want is a "one-stop, fast-responding" manufacturing partner. CanSemi Semiconductor's strategy is to deeply collaborate with industry-leading clients. Its customers include the world's largest fingerprint chip design company, the world's largest electronic tag chip design company, the world's second-largest image sensor chip design company by shipments, and two of the world's top three independent mobile phone chip design companies. By the end of 2025, the cumulative number of customers exceeded 200, with an 80% cooperation coverage rate among the top ten domestic listed analog chip companies. In the fourth quarter of 2025, its wafer shipments for high-voltage display driver chips ranked seventh globally and third in mainland China.
For a foundry established just nine years ago to make its name on global segmented rankings relies not on luck, but on the asymmetric competition logic of "specialty processes + deep collaboration."
Revenue Doubles in Three Years as Revenue Structure Quietly Shifts
The financial figures themselves are not astonishing: over the past three years, operating revenues were CNY 1.044 billion, CNY 1.681 billion, and CNY 2.582 billion, respectively, with a compound annual growth rate of 57.30% from 2023 to 2025. However, when breaking down the structure, this company is almost a cross-section for observing the transformation of China's chip industry.
Three clues are very clear. First, the process structure is moving upward. The revenue proportion of the 153nm/180nm nodes dropped from 49.14% to 33.22%, while the combined proportion of the 55nm/65nm, 90nm/95nm, and 110nm nodes rose from 24.91% to 45.45%, with a three-year compound growth rate of 111.88%. This also reflects that the products of Chinese design companies are moving towards the mid-to-high end, and the foundry side must keep pace.
Second, the downstream shift. The proportion of consumer electronics dropped from 88.75% to 77.79%, while revenue from industrial control increased from less than CNY 100 million to CNY 504 million, with the proportion breaking through 20%. Behind this is the massive wave of the entire Chinese chip industry migrating from consumer electronics to industrial and automotive sectors. The company has also passed the IATF 16949 automotive certification, built an automotive-grade quality system according to the VDA 6.3 standard, and clearly set a route of "consumer - industrial - automotive - AI (Artificial Intelligence)."
Third, the dual domestic and international circulation. The proportion of domestic revenue remains stable at 90%, while overseas revenue has achieved a compound annual growth rate of 111.57%, with customers spread across Taiwan, Hong Kong, and Singapore. During the window period of global industrial chain restructuring, mature node foundry services are becoming nodes for two-way flow.
The other side of the coin is equally evident: the proportion of the top five customers has risen to 62.68%, indicating increasing concentration; consumer electronics still account for nearly 80% of the revenue base; and the absolute amount of R&D investment has declined for three consecutive years, shrinking from CNY 605 million to CNY 422 million, with its proportion dropping to 16.36% as revenue scales up. These are real constraints, not fatal flaws, but they determine the height of the ceiling.
Silicon Photonics: A Rare First-Mover Advantage in the Global Coordinate System
If specialty processes are the survival logic of CanSemi Semiconductor, silicon photonics is its biggest highlight in the global industrial narrative.
At the end of 2024, it launched a 12-inch 90nm SiPho platform. As of April 2026, according to Frost & Sullivan data, it is the only enterprise in mainland China with the capability for large-scale mass production of 12-inch silicon photonics wafers, and the cumulative tape-out volume of its silicon photonics process technology platform has exceeded 7,000 wafers. Its products cover 400G, 800G, and 1.6T high-speed optical modules, which are on par with international leading levels. The 3.2T NPO product is currently being introduced, and subsequent R&D and mass production of 65nm silicon photonics and optoelectronic integration processes will be carried out, evolving towards NPO and CPO.
The weight of this strategic positioning becomes clear only from a global perspective: in the AI computing power competition today, the bottleneck is shifting from chip computing power to data interconnection between chips. The demand for optical interconnection is exploding, while the number of players globally capable of mass-producing 12-inch silicon photonics wafers is very limited. A Chinese foundry renowned for its mature nodes happens to be standing at the door of next-generation interconnection technology, which is a scarce asset in the global foundry industry.
Capacity expansion is another catch-up game. The existing two fabs are planned for 80,000 wafers per month, achieving 63,300 wafers by the end of 2025. The fourth phase plans to build an additional 40,000 wafers per month, with nodes scaling down to 65nm-22nm, serving platforms such as silicon photonics, embedded memory, image sensors, and computing-storage integration. After the completion of the fourth phase, the company's total capacity will reach 120,000 wafers per month. CanSemi Semiconductor candidly admits that its capacity scale is relatively small and economies of scale are hard to manifest. The fundraising is essentially a dual bet on capacity and processes. Among this, CNY 3.5 billion is planned for capacity expansion, accounting for about 46.7% of the total funds raised, nearly half; CNY 2.5 billion is allocated for cutting-edge R&D, making up 33.3% of the total, about one-third; and CNY 1.5 billion is used to supplement working capital, accounting for 20%.
Conclusion
Placing CanSemi Semiconductor back into the larger picture of the global semiconductor industry, its position is actually very clear: it is not to challenge the giants, but to be a professional player deeply rooted and fast-responding in the world of mature nodes and specialty processes; at the same time, leveraging the first-mover advantage in silicon photonics, it secures a scarce mass production position in mainland China in the global AI optical interconnection competition.
In nine years, from zero to CNY 2.5 billion in revenue, from zero wafers to a monthly capacity of 63,300 wafers, the growth of this company confirms a simple industrial logic: opportunities in semiconductors have never belonged solely to the most advanced nodes, but also to those who thoroughly understand segmented demands. Of course, customer concentration, reliance on consumer electronics, and the comprehensive gap with international giants are all obvious challenges. Listing is just raising the stakes, not the answer. Whether this "Guangzhou's First Chip" can grow into an indispensable piece in the global specialty process foundry landscape depends on the upcoming capacity ramp-up and product iteration, which will be the true test.
*Disclaimer: This article is created by the original author. The content represents their personal views. We repost it solely for sharing and discussion, which does not imply our endorsement or agreement. If you have any objections, please contact the backend.