At the beginning of August this year, Chuanhu (2059) jumped onto a server sliding rail worth over NT$10,000 per share, becoming the third “ten-thousand-share” stock on the Taiwan stock market, drawing market attention. In fact, the next candidate for a ten-thousand-share stock identified by foreign investors is Wanset (6223), a semiconductor test interface and equipment company. Although Wanset’s share price fell to the daily limit on August 17, closing at NT$5,800, analysts still remain optimistic about its future operating growth.
Aletheia Capital, an independent investment research firm in Hong Kong, raised its target price for Wanset from NT$4,000 to NT$10,000 in June this year. It also expects 2026 earnings per share (EPS) of at least 78.1 yuan, and projects that 2027 and 2028 will increase to 186.5 yuan and 286.5 yuan, respectively.
A local probe card manufacturer founded in Hsinchu in 1995—how did it manage to become the only Taiwanese company to break into the global top five, and even get foreign investors calling for a target price of 10,000? The key lies in the fact that AI chips are getting harder to test, making the probe card business increasingly valuable.
How important are probe cards as the bridge between wafers and testing equipment?
One of Wanjing’s core businesses is probe cards, which accounted for 72.2% of total revenue in 2025. Probe cards are indispensable consumables for wafer testing. As the complexity of AI chips increases, testing demand also rises.
After a wafer is made, it is not sent entirely to packaging right away. It first needs to confirm which chips on the wafer can be used and which have defects. A probe card is like the bridge between the wafer and the testing equipment: by using probes to make contact with the pads on the chip, it sends and receives electrical signals, completes testing, and only then proceeds to dicing and packaging.
Probe cards are mainly divided into three types: cantilever (CPC), vertical (VPC), and microelectromechanical systems (MEMS). The cantilever structure is simple and lower in cost, and is mostly used in mature processes. Vertical probes have higher probe density and are suitable for high-performance computing chips. MEMS probe cards use semiconductor process technology to fabricate probes, enabling finer pitch spacing, and are applied in advanced products such as AI chips and high-bandwidth memory (HBM). The structural differences among the three types of probe cards can be referenced in “What is a probe card? Which stocks are conceptually related? Illustrated guide to probe card structure and the features of three types of probe cards.”
A comparison of Wanjing, Yingswei, and Hongjin: the three probe card heroes
Wanjing, Yingswei, and Hongjin are often collectively referred to in the market as the “three probe card heroes,” but the truth is that the three companies occupy different positions in the semiconductor testing process.
Wanjing’s probe cards handle testing before wafer dicing. Yingswei (6515) focuses on coaxial test sockets used for post-packaging testing. Hongjin (7769) primarily handles testing sorters, which send large volumes of chips into testing equipment, categorize them based on test results, and control temperatures during the process.
Looking only at revenue scale, Hongjin is clearly larger than Wanjing, with 2025 revenue about 2.3 times that of Wanjing. However, the companies’ products differ, so you can’t directly judge competitiveness by revenue. The real reason investors favor Wanjing goes back to its core probe card business.
Wanjing (6223) Yingswei (6515) Hongjin (7769) Main products probe cards (CPC/VPC/MEMS), coaxial test sockets for testing equipment, vertical probe card testing sorters, active temperature-controlled systems, water-cooling plate testing flow location wafer-level (before dicing) package-level (after packaging) final test and system-level test 2025 revenue NT$13.371 billion NT$7.857 billion NT$30.271 billion 2025 gross margin 55.55% 45.26% 56.54% 2025 EPS NT$33.49 NT$46.93 NT$75.71
With demand exceeding supply, ramp up production all the way to 2027
First, look at the financials. In the second quarter, Wanjing revenue was NT$5.233 billion, up 33% quarter-over-quarter and 59% year-over-year. It delivered its sixth consecutive quarter of record highs, and quarterly revenue first surpassed NT$5 billion. After-tax net income attributable to the parent company was NT$1.524 billion, up 1.43 times year-over-year; EPS reached NT$15.55. In the first half, cumulative EPS was NT$28.08; after-tax net income attributable to the parent company was NT$2.751 billion, up 103.4% year-over-year.
More importantly, Wanjing’s available capacity is almost fully booked. Vertical and MEMS probe cards have been operating at full load since the second half of last year; cantilever cards are also fully loaded due to an influx of new orders. As a result, the company plans to expand production continuously from the second half of this year through 2027.
Even with orders filled, Wanjing did not choose to raise prices directly. Instead, it tends to negotiate with customers for advance payments and long-term demand commitments—customers willing to make such commitments are prioritized for capacity.
This effectively trades constrained near-term capacity for greater visibility into longer-term orders. Wanjing states that the visibility of its collaboration and development plans with customers is already beyond one year. This is also one of the reasons investors are optimistic about its profit growth going forward.
However, ramping up production itself also brings risks. What the company is betting on is demand after 2027. If AI capital expenditures slow down, the utilization rate of the new capacity may fall short of expectations. If North American CPU customers ramp up more than the market expects, customer concentration is also a metric to watch going forward.
“The only Taiwan company among the global top five”—how much room does Wanjing still have?
Wanjing chairman Ge Zhanglin said directly at the March 30 performance briefing at the OTC Market Center this year that Wanjing has “no competitors” in Taiwan.
Based on global rankings, this claim has its basis. According to market research firm QYResearch, the global probe card market size in 2024 was approximately USD 2.987 billion (about NT$95.3 billion). The market has long been led by US company FormFactor and Hong Kong? (actually: US? ) The ranking is led by FormFactor and Technoprobe, with Micronics Japan from Japan in the second place; Wanjing is the only Taiwanese supplier among the world’s top five providers.
Although there is still a clear gap in overall scale compared with the market leader FormFactor, Wanjing has already established a certain market position in cantilever and vertical probe cards. The real area where it still has to catch up is MEMS probe cards, which have higher technical barriers.
Wanjing only successfully developed and mass-produced MEMS probe cards in 2019. According to the company’s remarks at this year’s March earnings call, its market share is about 2%, and its global ranking is around 11 to 12. As demand for advanced testing is driven by AI chips and HBM, how much of this market Wanjing can capture will directly determine the size of its next growth phase.
From its founding in 1995 to today, Wanjing has not suddenly emerged just because of AI. Instead, it has built up more than 30 years in the probe card market. What AI has truly changed is that chips are becoming more and more complex, which in turn increases the importance of advanced testing.
Today, Wanjing is already among the world’s top five probe card makers. The next hurdle is the MEMS market it is still trying to catch up in. Whether the foreign investors’ shouted target price of 10,000 can come true ultimately depends on whether this round of capacity expansion can translate into orders—and how much of the advanced market Wanjing can capture again from the global leaders.
This article is reproduced with permission from: (Digitimes)
Original title: (The next Chuanhu? Wanjing’s chairman once said, “Taiwan has no competition.” Where does the confidence to call for a price of 10,000 come from?)
Original author: Lin Yuxuan
“The next Chuanhu? Foreign investors shout a target price of 10,000—how will Wanjing capture the AI chip testing opportunity?” The article was first published in “Crypto City.”
