Hangzhou Changchuan Technology: The Playbook of China's Semiconductor Test Champion

I. Introduction & Episode Roadmap (00:00:00 – 00:12:00)

Every chip ever shipped is interrogated before leaving the factory. Somewhere between a wafer coming off the fabrication line and a packaged IC landing in a smartphone, an electric vehicle, or an AI server rack, an automated system grips the die, presses electrical contacts against pads narrower than a human hair, and runs thousands of parametric tests in seconds: Does this pin drive the correct voltage? Does this memory cell hold its charge at 85 degrees Celsius? Does this RF front-end hit its target frequency?

Dies that pass are binned as good; those that fail are discarded. The machine running the test logic is a tester. The robotic system feeding it parts is a handler. The equipment conducting the same checks while dies remain on the wafer is a prober. For most of semiconductor industry history, virtually all test equipment deployed in China was imported from overseas suppliers.

This is the story of the company that set out to change that dynamic, and the capital-intensive path it followed. Listed as 300604.SZ on the 创业板 ChiNext board of the 深圳证券交易所 Shenzhen Stock Exchange, 杭州长川科技股份有限公司 Hangzhou Changchuan Technology Co., Ltd. began in 2008 as a spin-out from a Hangzhou integrated device manufacturer. It was founded by a former production director who brought key middle management with him.¹ Rather than attacking the industry's most complex equipment first, the company started with mechanical handlers and sorters, using cash flow and customer access from that beachhead to expand step-by-step up the value chain.

Why Test Equipment Is a Chokepoint

Understanding the market structure explains why industrial policy has focused heavily on a segment that, by one estimate, accounts for only about 6.7% of total semiconductor equipment value.² Automated test equipment functions as a critical industry bottleneck. Two companies — Teradyne of the United States and アドバンテスト Advantest of Japan — have historically controlled more than 80% of the global tester market, with their dominance most pronounced in advanced silicon.³ In high-performance computing and artificial intelligence segments, a research note following Changchuan's 2025 financial results estimated Advantest's market share at over 66%.⁴ Domestic test capabilities are therefore widely viewed as essential to broader semiconductor self-sufficiency.

The Thesis, Reduced to Two Numbers

The core investment thesis centers on whether Changchuan can transition from mature to advanced equipment categories. By 2024, Chinese localization rates in analog and discrete-device testers exceeded 80%. However, domestic market share stood at roughly 10% in SoC testers, approximately 8% in memory testers, and near 1% in radio-frequency testers.³ Changchuan's long-term valuation depends on expanding into these complex digital categories while maintaining profitable margins.

Recent financial disclosures indicate that this transition is underway. Revenue reached RMB 5.29 billion in 2025, a 45.31% increase year-over-year, while net profit attributable to shareholders rose 190.42% to RMB 1.33 billion.⁵ In the first quarter of 2026, revenue grew 69.09% and net profit increased 217.60%, results management attributed primarily to higher digital tester sales.⁵ Yet in 2023, the company reported a 31% revenue drop, net profit of just RMB 45 million, and an operating loss on an ex-items basis.⁶ Evaluating Changchuan requires reconciling this sharp cyclical volatility with its growth trajectory.

The Story Arc

Five core themes structure this story. First, the founding culture: what former engineers from 士兰微 Silan Microelectronics brought to the venture, and how close customer integration shaped early product development. Second, the platform strategy: a sequential expansion from handlers into analog testers, digital testers, and 12-inch wafer probers, with legacy lines funding higher-end R&D. Third, cross-border acquisitions: the 2019 purchase of Singapore-based Semiconductor Technologies & Instruments alongside subsequent expansion in Malaysia, comparing integration outcomes against management targets.

Fourth, the role of state capital: how the 国家集成电路产业投资基金 China National Integrated Circuit Industry Investment Fund (commonly known as the 大基金 Big Fund) acted as a key initial sponsor before systematically paring its equity stake. Fifth, corporate governance: capital allocation decisions, repeated secondary equity offerings, and historical insider selling trends.

The story begins in Hangzhou, inside an integrated device manufacturer that was itself born from an entrepreneurial spin-out.

↑ Back to Top

II. Founding Context & The "Silan School" (00:12:00 – 00:32:00)

In September 1997, eight founders departed a state-owned electronics enterprise in Shaoxing with pooled capital of RMB 3.5 million to form 士兰微 Silan Microelectronics. Among the young engineers who joined that July, fresh from a degree in semiconductor materials, was 赵轶 Zhao Yi. Over more than a decade at Silan, Zhao rose to production director — a factory role focused on operational uptime rather than academic research. His primary responsibility was keeping assembly and test lines running continuously.¹

A Factory Floor Education

That distinction shaped the company's early strategy. A production director at a Chinese integrated device manufacturer in the early 2000s operated under constant logistical friction: floor equipment was foreign, expensive, and supported at arm's length. When a handler jammed during a night shift, repairs depended on foreign vendor schedules, international currency billing, and multi-layered regional support desks. Zhao spent ten years analyzing which failure modes caused the greatest financial losses and where support bottlenecks occurred — an operational perspective distinct from university research environments.

In April 2008, Zhao departed to found the entity that became Changchuan. The name originated as a personal nod: Zhao's ancestral home is 四川 Sichuan, his wife and co-founder 徐昕 Xu Xin comes from 长兴 Changxing in Zhejiang, and the characters were combined as 长川 Changchuan.¹ A key structural factor was the talent that followed him. At least six Silan managers made the transition, including 孙峰 Sun Feng, 钟锋浩 Zhong Fenghao, 韩笑 Han Xiao, and 朱红军 Zhu Hongjun — engineers and department heads at Silan who subsequently became directors and vice presidents at Changchuan.¹ Several of these executives remain among the company's largest individual shareholders nearly two decades later.

The Import Gap of 2008

In 2008, domestic test equipment options were scarce. China's chip design firms and outsourced assembly and test providers — such as 长电科技 JCET, 通富微电 TFME, and 华天科技 Hua Tian Technology — were expanding rapidly. However, these providers relied almost entirely on imported back-end equipment from Teradyne, Advantest, 東京エレクトロン Tokyo Electron, and Cohu. While the customer base was domestic, the toolsets were entirely foreign. Entering this market presented substantial technical hurdles, as automated test equipment required specialized mechanical and electronic design capabilities.

Tester, Handler, Prober: What Each Machine Actually Does

Understanding Changchuan's entry strategy requires distinguishing the three core categories of back-end equipment:

A tester serves as an automated electrical evaluator: a rack of precision instrumentation that injects signals into an integrated circuit and measures the returned output under software control. Its development complexity lies primarily in high-speed electronics, signal integrity, and software architecture.

A handler is the robotic system that feeds packaged chips into the tester. It picks components from trays or tubes, conditions them to specified operating temperatures, presses them into contact sockets during testing, and sorts them into pass or fail bins at high speeds without damaging leads or packages. Its core engineering challenges involve high-precision motion control, thermal management, mechanical durability, and operational throughput.

A prober performs a similar mechanical positioning function for unpackaged dies while they remain on a semiconductor wafer, aligning fine probe needles onto die pads with micron-level accuracy across a motorized stage.

The Beachhead Decision

Zhao chose to enter the market via mechanical handlers rather than testers. Handlers offered structurally lower gross margins due to their reliance on mechanical assemblies rather than proprietary software and high-margin measurement electronics.

However, handlers established a critical operational beachhead. Installed directly alongside testers on the factory floor, handlers physically touched every production unit. Supplying handlers allowed Changchuan to participate in customer yield reviews, examine operational failure data, identify specific testing bottlenecks, and secure qualification as an approved equipment vendor. This vendor status created a distribution channel for subsequent product lines.

To complement this hardware beachhead, Changchuan adopted a localized service model. Situated in the Yangtze River Delta near China's main cluster of packaging and test facilities, the company competed on response times. While multinational vendors routed complex technical issues through overseas engineering teams, Changchuan dispatched application engineers locally on short notice. This strategy created an initial competitive buffer based on service responsiveness and proximity, allowing the company to build factory relationships while developing higher-complexity equipment.

In its initial phase, Changchuan functioned primarily as a specialized mechanical engineering firm leveraging local pricing and support. The central strategic question for its next growth phase was whether this operational foothold could support an expansion into high-margin electrical test equipment.

↑ Back to Top

III. The Handler Bridge: Scaling from Sorters to ATE (2008–2016) (00:32:00 – 00:55:00)

Picture a Chinese packaging plant around 2012. The factory operates across three shifts. A gravity-fed handler drops packaged components down a chute into a test socket, the tester fires, and the parts are sorted into bins. When the equipment functions smoothly, it attracts little notice. When a machine stops, an entire production line halts, bringing plant management to the floor within minutes. Changchuan's early commercial success was won and lost in those operational moments.

During its initial phase, the company scaled gravity-fed and pick-and-place handlers across domestic assembly and test facilities. The financial characteristics reflected the mechanical nature of the equipment. Handlers provided volume and operational cash flow rather than high profit margins—a dynamic still visible in recent performance, where handler gross margins remained in the low forties compared to the low sixties for tester lines.⁴ Management treated the handler segment as an operational bridge to secure customer access rather than a long-term destination.

That bridge led first into analog testing. Changchuan introduced test platforms targeted at analog and mixed-signal devices, including power management circuits and discrete components used in home appliances, chargers, motor controls, and automotive subsystems. Major incumbents such as Teradyne and Advantest focused the bulk of their R&D investments on digital SoC testers for high-end computing markets. Analog testing offered smaller overall market volume, but provided a less contested entry point where established vendors relied heavily on mature platforms.

Climbing the Tester Ladder

Entering the automated test equipment market encompasses widely varying levels of technical complexity. An analog tester evaluates continuous electrical parameters—such as voltage, current, and temperature coefficients—on components with limited pin counts and moderate signal speeds.

By contrast, a high-performance digital tester must drive and measure thousands of pins simultaneously at multi-gigabit speeds, keeping timing aligned within picoseconds while delivering stable power as internal chip blocks rapidly cycle on and off. Beyond hardware and high-speed signal integrity, software presents a formidable barrier. Pattern generators, compilers, and debug environments require years of user familiarity, creating substantial switching friction for customers accustomed to incumbent software stacks.

Changchuan expanded upward from lower-complexity entry points. The company introduced analog and mixed-signal platforms first, later adding higher-power analog systems such as the P3000 and multi-site mixed-signal architectures in the CTA8290 series before developing digital architectures.⁷ Commercializing hardware at each stage enabled the business to fund incremental R&D and build operational credibility before introducing higher-end equipment.

This expansion brought Changchuan into direct competition with 华峰测控 AccoTest, a Beijing-based vendor that had built a focused, high-margin position in analog and power testing. The two firms coexisted with sharply contrasting financial strategies.

In 2025, AccoTest generated RMB 1.35 billion in revenue with a company gross margin near 74% and R&D spending at 19.74% of sales.⁸ Over the same period, Changchuan generated nearly four times that revenue, operating at a lower blended gross margin in the mid-fifties but reinvesting more aggressively with R&D intensity above 23%.⁵⁴ Where AccoTest pursued high unit profitability in specialized segments, Changchuan sought to build a broader equipment platform across the entire test cell.

Why Qualification Is the Real Moat

Beyond mechanical scaling and product development, securing customer qualification represented the principal structural barrier during this period. Qualification creates substantial supplier stickiness because adopting a new test platform requires significant engineering investment from the customer.

A factory must write custom test programs, engineer dedicated interface boards to connect pin electronics to the device under test, correlate defect data against incumbent systems, and run parallel production trials over several months. The process involves significant engineering labor and operational risk, as passing defective chips carries severe commercial penalties. Test engineers rarely risk production yields solely to achieve marginal equipment cost savings.

Securing qualification at tier-one domestic packaging and test providers required years of continuous validation. Changchuan's core domestic customer base—comprising leading assembly and test providers alongside integrated device manufacturers like Silan—was established through this extended qualification cycle.⁷ High switching costs cut both ways: while long validation cycles delayed initial market entry, they subsequently protected installed positions against new rivals and made customer relationships difficult to displace.

By 2016, Changchuan had established a cash-generative mechanical handler business, an initial foothold in analog testing, a localized service infrastructure, and a portfolio of tier-one customer qualifications. However, the company lacked both the capital scale and advanced technology required to compete in high-margin digital tester categories. Addressing those gaps required external capital and strategic acquisitions, which arrived in 2017.

↑ Back to Top

IV. The 2017 SZSE Listing & Big Fund Alignment (00:55:00 – 01:15:00)

On 17 April 2017, Changchuan listed on the ChiNext board of the Shenzhen Stock Exchange — the first Chinese integrated-circuit packaging and test equipment maker to go public.⁹ The company that rang the bell was small by any standard: a niche equipment supplier with a domestic customer list and revenue in the low hundreds of millions of renminbi. The listing itself mattered less for the money raised than for what it made possible afterward — a currency. A listed equity gives an acquirer the ability to buy assets with paper, and it gives strategic investors a liquid instrument to hold. Changchuan used both, aggressively, within eighteen months.

Enter the Big Fund

The strategic investor arrived in December 2018. The Big Fund — the state vehicle created to accelerate China's semiconductor industry — participated in a private placement alongside 天堂硅谷 Tiantang Silicon Valley and a Shanghai equipment and materials fund, taking roughly 16.25 million new shares at RMB 30.17 apiece, about RMB 490 million in total, which lifted the Big Fund's holding from 7.28% to 9.85% and made it the second-largest shareholder.¹⁰ This was not a passive financial allocation. Changchuan became, in the language of the sell-side at the time, the only integrated-circuit test equipment platform the national fund had backed.

What did that actually buy? Three things, in descending order of durability. Capital, first, and on terms available to almost nobody else at that stage of the cycle. Validation, second: state sponsorship functioned as a signal to procurement committees at state-linked foundries and packaging houses that this vendor was strategically endorsed, which shortens the political half of a qualification decision even if it shortens none of the technical half. And third, introductions — access to the buyers whose equipment budgets were about to be redirected toward domestic suppliers under the broader 国产替代 domestic substitution push.

Discipline, or Empire-Building?

The critical question for a long-term investor is what the company did with the money, and here the record is genuinely impressive on one axis and genuinely uncomfortable on another. On the impressive side, research intensity ratcheted up and stayed up. R&D investment reached RMB 1.02 billion in 2024, equal to 28.14% of revenue, and RMB 1.27 billion in 2025, or 23.97% of a much larger revenue base.¹¹⁵ The ratio falling while the absolute number rose is what operating leverage looks like in an R&D-heavy business, and it is the single clearest quantitative signal that the earlier spending was finally converting into sellable product.

Research staff numbered 2,059 as of mid-2025, more than half the company's roughly three thousand employees.¹¹ For a capital equipment maker, a workforce that is half engineers is itself a strategic statement: it is a cost structure that only pays off if the technology gap actually closes.

There is a second measure of whether the spending was disciplined or merely large, and it is the accumulation of defensible intellectual property. By early 2025 the company had built a portfolio of 933 authorized patents — 328 of them invention patents, the category that requires substantive novelty examination rather than mere utility — plus 545 utility patents and 72 software copyrights, with R&D spending compounding at roughly 50% a year between 2020 and 2024.⁷ Patent counts are a crude proxy and easily gamed, but the mix matters: a portfolio weighted toward invention patents and software copyrights is more consistent with genuine architecture development than with incremental mechanical tweaks.

Combined with the segment evidence that arrived later, the fairest reading is that the post-listing capital went into real technology rather than into empire-building. Notably, Changchuan did not diversify. There is no adjacent-industry adventure in this history, no property arm, no unrelated acquisition spree — the classic diworsification failure mode of cash-rich Chinese listed companies is simply absent from the record.

On the uncomfortable side, this has been an equity-hungry company. The Big Fund placement was the first of several trips to the market, and the pattern has continued into the present. Zhao Yi's controlling position has been steadily worn down. As of the 2025 report he held roughly 22.3% directly, with Xu Xin's investment partnership holding a further stake for combined control near 28.5%.¹²¹³ For a founder who started with an unlisted company he wholly controlled, that is a long slide — and, as the next sections show, not all of it was caused by issuing shares to buy assets.

The strategic logic behind the spending is easier to defend than the financing method. The money went where the barriers are highest: 12-inch fully automatic wafer probe stations, and digital tester architectures. Both are multi-year, capital-destroying projects with no revenue for years and a real chance of failure. A company that wanted to optimize near-term earnings would not have made those bets. A company that wanted to be more than a handler supplier had no alternative. Whether that judgment was correct is now, finally, becoming measurable — but before the payoff arrived, the company took a detour through Singapore.

↑ Back to Top

V. M&A Strategy: The STI Singapore Acquisition Benchmark (01:15:00 – 01:42:00)

Certain specialized capabilities cannot be easily built from scratch; machine vision for semiconductor inspection is one of them. What sounds like a simple task—photographing a chip to detect defects—presents severe technical hurdles. Inspection algorithms must distinguish a fatal crack from a harmless surface mark at high production speeds across thousands of package types, keeping false-reject rates low enough to prevent discarding viable chips. Developing that capability requires decades of accumulated defect libraries and field tuning that cannot be rapidly replicated on a five-year timeline.

To acquire these tools, Changchuan turned to cross-border M&A. In late 2018, the company announced an equity-funded transaction to acquire 90% of 长新投资 Changxin Investment, a holding entity that owned 100% of Singapore-based Semiconductor Technologies & Instruments (STI). Combined with a 10% stake acquired earlier that year through a direct capital injection, the transaction gave Changchuan full ownership of STI.¹⁰ The deal valued STI at approximately RMB 545 million, with roughly RMB 490 million paid through newly issued shares for the final 90% tranche.¹⁰ The acquisition closed in September 2019, and the new shares were listed later that month.

STI was an established industry player. Operating for over two decades as a subsidiary of Singapore-listed ASTI, the firm held more than 150 patents and manufactured turret handlers, optical inspection systems, and wafer-level automated optical inspection tools. Crucially, STI brought an international customer base that Changchuan could not have quickly assembled domestically, including Texas Instruments, Micron, STMicroelectronics, Samsung, and leading assembly-and-test providers ASE and Amkor. STI had previously been recognized as one of Texas Instruments' top fifty global equipment suppliers.¹⁴

The Exchange Asks a Question

However, the transaction structure raised regulatory scrutiny. In December 2018, the Shenzhen Stock Exchange issued an inquiry letter asking why the deal lacked standard performance guarantees, profit compensation mechanisms, or earn-out price adjustments from the sellers of Changxin Investment. Regulators questioned whether the arrangement adequately protected Changchuan and its public shareholders.

A profit condition did exist deeper in the transaction chain: of the approximately SGD 90 million paid for STI, SGD 9 million was placed in escrow, contingent on STI achieving at least SGD 17 million in combined pre-tax profit across 2018 and 2019. However, that escrow protected the intermediate holding entity rather than Changchuan's public shareholders directly. The exchange's inquiry effectively centered on who would bear the financial risk if STI underperformed.

Public filings did not clarify whether the escrowed funds were ultimately released or clawed back after the global memory and packaging market entered a cyclical downturn shortly after closing. Nevertheless, STI recovered rapidly. In the first half of 2020, STI generated RMB 171 million in revenue—a 26% year-over-year increase—while net profit jumped 190%. By the first half of 2021, the Singapore unit contributed approximately 42% of Changchuan's total net profit.¹⁴

For minority shareholders, the structural takeaway outweighed the short-term numbers. Changchuan executed a major cross-border acquisition using equity financing, while placing earnings protections one corporate layer removed from the parent company's shareholders. Although STI ultimately delivered strong earnings, reliance on paper-funded acquisitions with indirect minority safeguards became a recurring pattern in Changchuan's capital allocation strategy.

The strategic benefits, however, proved substantial. Integrating STI's optical inspection capabilities enabled Changchuan to offer combined electrical testing, mechanical handling, and visual inspection within a single integrated test cell—a combination unmatched by domestic peers. STI's high-precision two-dimensional and three-dimensional inspection tools remain deployed across major global semiconductor accounts.⁴

The acquisition also expanded Changchuan's geographic footprint. The transaction provided established service teams and subsidiaries across Southeast Asia, South Korea, and Europe, along with direct commercial access to major tier-one chipmakers. Overseas sales expanded to represent roughly one-quarter of Changchuan's total revenue, delivering higher unit pricing and stronger gross margins. Identical equipment commands prices over 10% higher in international markets, generating gross margins several percentage points above domestic sales.²² Furthermore, an established Singapore-based operational footprint provided a hedge against expanding global trade restrictions.

EXIS, Kewei, and the Buy-What-You-Cannot-Build Rule

Changchuan subsequently repeated this targeted M&A playbook. The company acquired Malaysia-based EXIS Tech through an equity-funded deal approved in 2022 and completed in the first half of 2023. EXIS added high-speed turret handlers—spanning the 250 through 700 series with 20 to 40 test stations—allowing Changchuan to offer a complete handler portfolio across gravity, pick-and-place, and turret architectures.⁷ Changchuan then acquired vision software specialist 科为升 Kewei Vision in a two-step transaction: purchasing a 51% controlling stake for RMB 96.9 million in December 2024 at a RMB 190 million valuation, and agreeing to buy the remaining 49% for RMB 119 million in September 2025 at a RMB 243 million valuation, supported by a merger loan of up to RMB 60 million from the Export-Import Bank of China.¹⁵

Taken together, Changchuan's acquisitions reflect a consistent strategic focus: acquiring specialized technical capabilities—such as optical vision algorithms and complex turret mechanics—that would require years to develop internally, rather than purchasing sheer scale or market share. While individual deal sizes remained modest relative to Changchuan's overall balance sheet, reliance on equity issuances created ongoing shareholder dilution. This trade-off between technology acquisition and equity expansion would reappear on a larger scale in subsequent capital raises.

Before examining those larger capital moves, it is necessary to examine the high-end digital test market Changchuan set out to challenge.

↑ Back to Top

VI. Industry Structure & Competitive Dynamics: ATE & Equipment (01:42:00 – 02:15:00)

If you wanted to design an industry structure hostile to new entrants, you would design the automated test equipment market. Two firms — Teradyne and Advantest — have jointly held over 80% of the global tester market, and their positions are strongest precisely where the chips are hardest.³ Their advantage compounds in a specific, brutal way: the leading-edge customers who need the newest test capability co-develop it with them, years before the chips exist. By the time a challenger can buy the same components off the shelf, the incumbent has already shipped, been qualified, and moved on.

The AI cycle has made this worse, not better. Advantest's share of the SoC tester segment reportedly climbed from the mid-fifties in calendar 2024 to roughly two-thirds in calendar 2025, and one research note published after Changchuan's 2025 results put Advantest above 66% of the global test equipment market.⁴ Both incumbents publish their results in detail, and their investor materials describe a business enjoying the strongest demand conditions in its history, driven by compute and high-bandwidth memory.¹⁶¹⁷ The relevant point for a Changchuan investor is uncomfortable: the incumbents are not weakening. They are getting stronger in the very segments Changchuan says it is entering.

Against that, the China market has been carved open by policy rather than by technology. The localization picture as of 2024 was starkly bifurcated. Analog testers: above 80% domestic. Discrete-device testers: above 80% domestic. SoC testers: about 10%. Memory testers: about 8%. RF testers: about 1%.³ And the size distribution is inverted relative to the localization progress — SoC testers alone accounted for 63.14% of China's tester market in 2024, with memory testers a further 15.95%.³ Chinese equipment makers had won most of the market they could reach, and almost none of the market that mattered.

This is the whole investment case in one paragraph. The remaining opportunity is not incremental. It is the majority of the market, and it is currently held almost entirely by two foreign firms whose products are better.

The Domestic Map

Within China, the competitive map divides cleanly. In analog and mixed-signal test, AccoTest is the entrenched specialist, running a leaner and far more profitable model on a narrower product range. In handlers and optical inspection, Changchuan holds a leading domestic position alongside foreign incumbents including Cohu and Besi, materially helped by the STI and EXIS acquisitions. In probers, the historic duopoly has been Tokyo Electron and 東京精密 Tokyo Seimitsu, and domestic penetration has been minimal — which is why Changchuan's 12-inch program has been treated as a national priority rather than merely a product roadmap item.

The optical inspection niche deserves its own note, because it is the one place where Changchuan competes internationally from a position of established strength rather than as a challenger. The Singapore product line it acquired was already selling to Texas Instruments, Micron, STMicroelectronics, Samsung and the large assembly houses before Changchuan owned it.¹⁴ That is an unusual asset for a Chinese equipment company — a franchise qualified at Western and Korean customers whose purchasing decisions are not influenced by Chinese industrial policy in any direction. The honest caveat is that it is also the smallest of the three businesses, so its strategic value exceeds its current financial weight.

The Prober Problem, in Plain English

That prober program is worth a moment of plain English. A wafer prober must position an array of needles onto pads a few tens of microns across, on a 300-millimeter wafer that may be slightly warped, while the whole assembly is held at a controlled temperature and moved thousands of times without drift. It is a precision-mechanics problem of the same family as a machine tool, with the tolerances of a microscope.

The current status, disclosed through investor communications in late 2025, is specific and useful: the CP12 variants for SoC, CMOS image sensors and discrete devices have reached mass-production status, while CP12-Memory remains in customer certification and small-batch promotion, with the company citing technical difficulty and long certification cycles and extending the project timeline.¹⁸¹⁹ Management also flagged the next round of technical upgrades it is pursuing — overhead rail systems, active temperature control, warped-wafer handling.¹⁹

That disclosure deserves credit for specificity. A company inclined to obscure would have said "probe stations are progressing well." Saying which variants are in production and which are not, and admitting a timeline extension, is the kind of granularity that lets outsiders keep score.

The Digital Battleground

The digital tester battleground is where the story either becomes historic or becomes ordinary. Changchuan's D9000 series is positioned as covering SoC, logic, mixed-signal and RF devices, for both wafer-level and final test, with the company claiming performance that leads domestically and approaches — the careful word is approaches — advanced international levels.⁴² The demand pull is real: China's domestic AI accelerator programs, automotive chips and communications silicon all need test capacity, and every one of those chips is harder to test than the analog parts Changchuan cut its teeth on. Digital testers are the product line management named specifically when explaining the 2026 acceleration.²⁰

The evidence that would settle the argument is adoption at premier domestic fabless AI design houses, in volume, at prices that hold. That evidence is accumulating in the revenue line but has not been disclosed customer by customer. Until it is, the digital tester claim remains a strong inference from segment growth rather than a verified fact.

↑ Back to Top

VII. Segment Breakdown, Financial Engine & Materiality (02:15:00 – 02:40:00)

The 2025 annual report is the first document in Changchuan's history where the platform strategy shows up unambiguously in the arithmetic.¹² Start with the shape of the business.

Testers generated RMB 3.20 billion in 2025, up 55.29%, and accounted for 60.53% of total revenue, at a gross margin of 62.25%.⁴ This is the profit engine, and its growth outpacing the company average is the single most important fact in the report — it means mix is shifting toward the high-margin end without the company having to shrink anything. The gross margin declined 4.36 percentage points year over year, which deserves a note rather than alarm: a tester line ramping new product families into new customers typically prices to win the qualification, and early units carry immature costs.

Handlers contributed RMB 1.57 billion, up 31.76%, at a gross margin of 41.34% — and here the margin moved the other way, improving by 4.99 percentage points.⁴ That is the more subtle signal. Handlers are the mature, competitive, mechanically-driven end of the portfolio, exactly where price wars are supposed to happen. Margin expanding in that line suggests either genuine manufacturing scale benefits or a mix shift toward the more sophisticated turret machines acquired with EXIS. Either way, the feared commoditization of the volume business did not show up in 2025.

Everything else — probers, optical inspection, parts and service — contributed roughly RMB 520 million, up 33.86%.⁴ Sized honestly, this is still the small end of the business, but it is the strategic option. Advanced packaging is the reason. As chips are increasingly built by stacking and stitching multiple dies together — chiplets, 2.5D and 3D structures — the economics of test change fundamentally. In a monolithic chip, one bad die is one bad chip.

In a stacked package, one bad die ruins every good die stacked with it. That makes testing at the wafer level, before assembly, dramatically more valuable, and it multiplies the amount of optical inspection required at every bonding step. One analysis of high-bandwidth memory noted that stacking pushes the number of required test insertions from three or four to more than fifteen, adding entirely new steps for die screening, known-good-stacked-die testing and post-stacking retest.²¹

The Whipsaw

The full company arc puts the 2025 result in perspective. Revenue was RMB 2.58 billion in 2022, then collapsed 31.1% to RMB 1.78 billion in 2023, with net profit falling roughly 90% to RMB 45 million and an ex-items loss of RMB 76.56 million.⁶ Then RMB 3.64 billion in 2024, up 105.2%, with net profit of RMB 458 million, up 915.1%, and operating cash flow of RMB 626 million.¹¹ Then the 2025 figures above. Then a first quarter of 2026 with revenue of RMB 1.38 billion and net profit of RMB 353 million, at a gross margin of 56.81%.⁵⁴

Look at the whipsaw rather than the level. A 31% revenue decline produced a 90% profit decline. This is what high operating leverage does in both directions, and it is the defining financial characteristic of the business. Roughly a quarter of revenue goes to R&D and it does not flex with the cycle, so profit is a small residual sitting on top of a large fixed cost base. Investors extrapolating the 2025–2026 profit growth rate should hold the 2023 experience in the other hand.

The quality of the recent growth is better than the headline, in one specific respect. In the first quarter of 2026 ex-items profit grew faster than reported profit, which in turn grew faster than revenue — meaning the acceleration came from core operations rather than subsidies or investment gains, and that core net margin expanded from the mid-single digits to the low twenties in a single year.²² That is what operating leverage looks like when it finally turns.

Returns on capital in this business are not a stable series and should not be modelled as one. The denominator moves in steps: capacity, overseas facilities and working capital all expand ahead of the revenue they are meant to serve. Total assets grew 23.0% in 2024 to RMB 7.26 billion while shareholders' equity rose 14.4% to RMB 3.29 billion, meaning the balance sheet was expanding faster than the equity base even before the current ramp.¹¹ In a peak year, returns on that capital look excellent; in a trough year, the same asset base sits idle and returns approach zero. Any single-year return metric for this company describes where it sits in the cycle more than it describes the quality of the business.

Where the Caution Lives

The balance sheet is where the caution lives. Inventory rose 59.22% in 2025 to RMB 3.56 billion, contract liabilities more than tripled to RMB 1.00 billion, and operating cash flow fell 10.10% to RMB 562 million.⁴ Then in the first quarter of 2026, operating cash flow was negative RMB 209 million, attributed to a surge in materials purchases.⁴ Independent analysis of that quarter flagged receivables equal to about 154% of quarterly revenue, inventory at 34.9% of total assets, and impairment charges of RMB 58 million against RMB 2.87 million a year earlier.²²

Two readings compete, and both are legitimate. Rising inventory and contract liabilities ahead of a delivery ramp are exactly what a capital equipment maker with a growing order book looks like — contract liabilities are customer prepayments, and tripling them is a demand signal, not a warning. But the ratio of operating cash flow to net profit at roughly 0.42 times in 2025 means less than half the reported earnings converted to cash.²² For a company reporting triple-digit profit growth, that gap is the number worth watching.

Equipment businesses with long build and acceptance cycles legitimately run negative working capital swings during ramps. They are also the businesses where revenue recognition on multi-stage acceptance can flatter a period. Nothing in the disclosed record suggests the latter. But the cash conversion needs to normalize, and 2026 is the year it should.

↑ Back to Top

VIII. Management Credibility, Governance & Capital Allocation (02:40:00 – 03:00:00)

Evaluating management requires examining capital allocation when short-term earnings pressure suggests pulling back. On that measure, Changchuan's record splits sharply between operational decisions, which reflect consistency and discipline, and shareholder-facing actions, which present distinct governance questions.

On the operational side, management maintained its research commitment through a sharp industry downturn. During the 2023 revenue contraction, the company kept R&D spending elevated at RMB 790 million against revenue of RMB 1.78 billion, choosing not to cut research to inflate short-term operating profit.⁶ It retained a research organization of over two thousand personnel through a period of minimal earnings. The operational acceleration in 2024 and 2025 followed directly from this sustained outlay, offering concrete evidence of an engineering-focused strategy rather than simple policy dependency.

Management's technical communications have also proven unusually specific for a domestic equipment provider. The company disclosed which wafer prober variants had reached mass production and which remained in customer certification, acknowledged timeline extensions with explicit rationale, and identified specific upgrades on its product roadmap.¹⁸¹⁹ Similarly, filings framed the early 2026 growth acceleration consistently across the first-quarter report and half-year pre-announcement, attributing momentum to prior R&D outlays converting into commercial tools, expanding downstream digital demand, and operating leverage expanding margins.²⁰

Corporate communications differ structurally from Western standards. Changchuan does not host quarterly earnings calls featuring open analyst questioning. Instead, management updates surface through the Shenzhen exchange's investor interactive platform, periodic briefing sessions, and broker channel checks. Institutional focus remains legible through the specific queries management addresses: the manufacturing status of individual prober models, digital tester wins at tier-one fabless design houses, margin recovery trajectories, and supply chain security.

The disclosures display a distinct pattern. On technical roadmaps and validation timelines, management provided concrete details, naming specific product variants and detailing schedule adjustments.¹⁸¹⁹ Regarding cyclical swings, management remained cautious, linking performance to earlier R&D investments rather than issuing quarterly forecasts.²⁰ However, disclosures omit the single metric required to evaluate its high-end transition: a granular revenue breakdown isolating digital and memory tester sales from mature analog platforms. Without this transparency, investors must infer the progress of high-end commercialization from aggregate segment growth.

The Selling

Governance considerations present a sharper contrast, marked by persistent insider equity sales. A widely cited 2024 analysis calculated cumulative insider share sales at approximately RMB 3 billion, encompassing founder Zhao Yi, his family investment vehicle, executive co-founders from Silan Microelectronics, and the Big Fund.⁶ The Big Fund's divestment trajectory was notable: after peaking above 10%, the state fund reduced its holding below 5% by the first quarter of 2024, generating roughly RMB 1.2 billion in proceeds.⁶ While Phase I of the Big Fund operated with a pre-defined capital return timeline, its exit removed a major institutional sponsor from the order book.

Founder-level share sales pose more direct questions for public investors. In November 2023, Changchuan Investment—a holding vehicle controlled by co-founder Xu Xin—proposed a block transfer of 5% of the company at RMB 29.60 per share, representing a 25% discount to the prevailing market price.⁶ A second proposed 5% transfer in 2025 was subsequently terminated on 27 August 2025.¹² Between late August and 24 September 2025, the partnership instead liquidated a 1% stake—comprising 11,219,977 shares—in the open market.¹² On 28 November 2025, management disclosed an additional divestment plan to sell up to 11,968,400 shares, or 1.8866% of equity, for estimated proceeds of RMB 923 million between 22 December 2025 and 21 March 2026, citing external capital allocation needs. Prior to this plan, cumulative sales by the entity since December 2020 totaled 16.16 million shares for approximately RMB 925 million, reducing its remaining holding to 4.15%.¹³

These recurring transactions create an obvious tension: while the company consistently seeks external equity capital to fund R&D against international incumbents, entities linked to founding leadership have systematically monetized personal shareholdings. Although common among growth companies on domestic exchanges, these simultaneous cash-outs and capital calls create an ongoing overhang for minority shareholders.

The Dilution Arithmetic

The scope of shareholder dilution became explicit in mid-2025. In June 2025, Changchuan announced a private placement proposing to issue up to 189 million new shares to support an R&D expansion program budgeted at RMB 3.84 billion.²³ Updated filings detailed target net proceeds of up to RMB 3.127 billion, allocating RMB 2.19 billion to automated test equipment and optical inspection development across subsidiaries in Harbin, Shanghai, and Suzhou over five years, alongside RMB 934.6 million for working capital.²⁴ The Shenzhen Stock Exchange review committee approved the application, with final registration before the China Securities Regulatory Commission pending in early 2026.²⁴ If fully executed alongside earlier transfers, the founding couple's combined equity interest would decline from 28.48% to approximately 18.10%, though management affirmed their voting control would remain intact.²³

This equity expansion can be interpreted in two ways. From a strategic perspective, absorbing dilution is an unavoidable prerequisite for challenging entrenched global equipment incumbents, where scaling R&D requires continuous capital deployment.

From a corporate governance perspective, combining institutional dilution with personal insider sales transfers execution risk to minority shareholders. Throughout its growth phase, Changchuan has relied heavily on equity issuances to fund technology acquisitions and product development, requiring public shareholders to absorb both share dilution and cyclical earnings volatility.

Accounting Judgments Worth Watching

Financial reporting practices present additional items for ongoing scrutiny. In 2025, total R&D outlays reached RMB 1.27 billion, whereas recognized R&D expenses on the income statement totaled RMB 936 million.⁴ The resulting difference indicates that substantial development costs were capitalized onto the balance sheet as intangible assets rather than expensed immediately. While capitalizing development costs is standard practice under domestic accounting standards once projects meet technical feasibility milestones, it defers cost recognition to future periods.

If capitalized R&D tools encounter commercial delays or adoption hurdles during customer certification, those intangible assets face potential future impairment charges. Additionally, financial governance records reflect turnover in senior accounting roles, with two chief financial officers departing between the 2017 listing and 2024, representing an average tenure of roughly two years.⁶ For institutional investors, tracking capitalized asset amortization alongside executive stability remains essential to evaluating the sustainability of reported earnings.

↑ Back to Top

IX. Strategic Frameworks: 7 Powers & Porter's 5 Forces (03:00:00 – 03:20:00)

Strip away the national policy narrative and ask the primary structural question: what would prevent a competitor from capturing this market share over the coming decade?

Under Hamilton Helmer's 7 Powers framework, the company's strongest claim rests on switching costs, an advantage that is genuinely robust. The qualification process detailed earlier — developing test programs, engineering custom interface boards, conducting correlation runs, and executing parallel production trials — imposes substantial engineering costs and operational yield risks on customers seeking to switch vendors.

Once Changchuan's software architecture and hardware interfaces are integrated into a packaging facility's workflow, the incumbent advantage shifts in its favor. The clearest empirical evidence of this moat was the company's resilience during the 2023 industry downturn: while revenue dropped by nearly a third, the customer base remained intact, enabling a rapid financial recovery in 2024. Cyclical volume contractions differ fundamentally from competitive churn, and distinguishing between the two is essential to evaluating the business.

Counter-positioning is moderate and steadily decaying. Changchuan's initial competitive edge rested on a responsive service and pricing model that foreign multinationals could not match without undercutting their own higher-margin economics — deploying application engineers locally on short notice, offering customized modifications, and maintaining domestic pricing structures. While effective against foreign vendors operating through regional support desks, this strategy provides little differentiation against domestic peers such as AccoTest or newer entrants operating with identical cost structures. Counter-positioning deters incumbents constrained by legacy economics; it offers little protection against domestic rivals using the same operational playbook.

Process power represents a compelling yet unproven capability at scale. The company's accumulated engineering know-how in high-throughput thermal handling and sub-micron positioning for 12-inch probers constitutes tacit organizational knowledge that cannot be rapidly duplicated. Achieving mass-production status on CP12 variants for SoC, CMOS image sensors, and discrete devices demonstrates tangible progress.¹⁸ However, with the CP12-Memory prober remaining in customer certification, process power in the highest-complexity equipment categories remains emerging rather than fully established.¹⁹

Scale economies represent the newest strategic power, becoming visible in the 2025 financial expansion. Spreading more than RMB 1 billion in annual R&D expenses across a revenue base that roughly tripled over three years explains the sharp expansion in net operating margins. Yet scale remains a relative metric: global leaders continue to spend multiples of Changchuan's total R&D budget annually. Consequently, Changchuan possesses scale relative to domestic competitors, but it has not yet achieved parity with the global incumbents it seeks to displace.

Two powers remain conspicuously absent. There is no branding power in industrial capital equipment, where purchasing decisions depend on rigorous technical evaluation and yield metrics rather than brand perception. Nor are there network economies, as an automated tester gains no intrinsic utility simply because other semiconductor manufacturers adopt the same platform.

Porter: Three Favorable Forces, Two Hostile

Applying Porter's Five Forces framework reveals an industry structure that is favorable in three dimensions and hostile in two.

Threat of substitutes is very low, representing the most durable structural feature of the automated test equipment sector. Integrated circuits cannot ship without electrical and parametric validation, and rising device complexity increases total test intensity over time. Equipment providers supplying essential quality-assurance tools to zero-defect manufacturing environments occupy a secure position within the supply chain.

Threat of new entrants is low under standard market conditions, constrained by multi-year technical development timelines, substantial R&D capital demands, and stringent customer qualification protocols. However, industrial policy initiatives have temporarily lowered effective entry barriers within China, where capital availability and domestic procurement preferences encourage foundries and test houses to evaluate new domestic suppliers.

Bargaining power of buyers is high. China's packaging and test sector is concentrated among a small group of major outsourced assembly and test providers and integrated device manufacturers. Equipment capital expenditure is cyclical and discretionary. During industry downturns, buyers defer equipment purchases — as occurred in 2023 — while during recoveries, they leverage competing domestic bids to negotiate favorable pricing.

Bargaining power of suppliers is moderate, representing a key strategic vulnerability. The performance of high-end testers depends heavily on specialized external components that Changchuan does not manufacture internally, including high-speed data converters, programmable logic devices, precision linear motors, and optical sensors. Historically, many of these critical sub-assemblies have originated from overseas suppliers in the United States and Japan.

Rivalry among existing competitors is high and intensifying. Domestically, Changchuan competes against a more specialized, high-margin peer in analog testing, alongside emerging domestic startups and buyers who actively pit vendors against one another. Internationally, it confronts two global incumbents whose market share in advanced digital segments has expanded further during the artificial intelligence expansion cycle.⁴

In summary, while the industry structure remains favorable and Changchuan's switching-cost moat is well demonstrated, its scale advantage remains largely domestic. The structural attributes required to transform the company from a national champion into a globally competitive leader — namely, leading-edge process power and R&D scale comparable to Teradyne and Advantest — remain work in progress.

↑ Back to Top

X. Skeptical Investor Stress Test & Current Risk Radar (03:20:00 – 03:30:00)

Take the bull case as read and go looking for the ways it breaks. Start by clearing away three pieces of consensus that do not survive contact with the filings.

Myth: Changchuan is China's answer to Teradyne. Reality: it is, today, primarily a handler and mid-range tester company that has begun selling higher-end testers. Testers of all kinds represented 60.53% of 2025 revenue, and the company has not disclosed how much of that came from digital SoC and memory platforms as opposed to the analog and mixed-signal machines it has sold for years.¹ The comparison that fits the current business is not with a leading-edge tester giant but with a broad back-end equipment supplier — closer in shape to Cohu than to Advantest.

Myth: state backing means guaranteed demand. Reality: the Big Fund has been a seller, not a buyer, for years, cutting its stake below 5% by early 2024.⁶ Policy has opened doors; it has not signed purchase orders. And the same policy that pushes a domestic foundry toward Changchuan pushes it toward every other Chinese equipment startup, several of which are now well capitalized.

Myth: the 2023 collapse was a company-specific stumble. Reality: it was the cycle doing what the cycle does, and it will happen again. Nothing in the intervening three years has reduced the operating leverage that turned a 31% revenue decline into a 90% profit decline.⁶

With the narrative cleared, the genuine risks are these.

The Risks That Are Real

The component chokepoint is the risk that could do the most damage fastest. A digital tester's channel card is built around high-speed analog-to-digital and digital-to-analog converters, precision timing circuits, and programmable logic — the same categories of components that sit squarely inside the scope of Western export control regimes. If Changchuan's high-end platforms depend on foreign-sourced parts in these categories, then a tightening of controls does not merely slow the company down; it halts specific product lines. The company has not published a component-level localization disclosure, so an outside investor cannot verify the exposure. That absence is itself the finding: the single largest identified risk to the highest-value product line cannot currently be quantified from public information. Research notes covering the stock routinely list geopolitical change as a primary risk without being able to size it either.²¹

The advanced-node reality check is the risk that matters most to the valuation. The skeptic's thesis is straightforward: Changchuan has won where the technology is easy and foreign vendors did not fight, and localization statistics support exactly that reading — above 80% in analog and discrete, around 10% in SoC, around 8% in memory, and roughly 1% in RF.³ The company's answer is its digital tester ramp, and the numbers are moving: testers grew 55% in 2025, the fastest of any segment, and management attributed the 2026 acceleration specifically to digital testers.⁵²⁰ What is missing is granularity. The company has not disclosed digital tester revenue as a separate line, nor named the leading-edge design houses that have qualified the platform. Segment growth is consistent with the bull case, but it is not proof of it. An investor seeking proof should demand a disclosed digital tester revenue split before treating the leading-edge claim as established.

Pricing pressure is a live and asymmetric risk. The margin evidence in 2025 was mixed in an informative way: handler margins rose while tester margins fell 4.36 percentage points.⁴ The comfortable explanation for the tester decline is new-product ramp costs and introductory pricing to win qualifications. The uncomfortable explanation is that domestic rivals are catching up on lower-difficulty tester families and competing on price. Both mechanisms produce the same reported number, and 2026's margin trajectory is what separates them — with the first quarter's 56.81% blended margin, up 4.06 percentage points year-over-year, providing early evidence for the benign reading.⁴

Customer capex cyclicality is the risk that is certain to recur. The only question is when. Equipment orders are a derivative of customer utilization, and utilization turns fast. The 2023 experience — revenue down 31%, profit down 90% — serves as the model, and nothing about the business has changed to soften it. If anything, a higher fixed cost base makes the next downturn's profit decline steeper. Any investor holding this business through a cycle should expect at least one year where earnings functionally disappear.

Working capital is the near-term risk that is already visible. Negative operating cash flow in the first quarter of 2026, receivables running well above a quarter's revenue, inventory approaching 35% of total assets, and impairment charges up twentyfold year-over-year are all consistent with a fast ramp — and all consistent with the early stages of a collection problem.²² The distinguishing evidence will show up in the cash flow statement over subsequent quarters, not in the income statement.

Customer concentration sits underneath the receivables risk. China's assembly-and-test industry is dominated by a small number of very large buyers, which means a receivable book running well above a quarter's revenue is not a diversified credit exposure, but a concentrated one.²² The same concentration that makes qualification valuable — win one of these customers and the volume is enormous — makes the collection profile lumpy and the negotiating position weak. This is a structural feature of selling capital equipment into an oligopsony, not a sign of mismanagement, but it belongs in any assessment of balance-sheet quality.

Finally, the funding overhang. The RMB 3.13 billion placement remained pending CSRC registration as of early 2026.²⁴ If it completes, dilution is the cost. If it does not, the R&D program either slows or gets funded from a balance sheet that is already absorbing a large working capital build. Neither outcome is catastrophic, but both are worth pricing.

↑ Back to Top

XI. Investment Story Spine: Bull vs. Bear & Playbook (03:30:00 – 03:40:00)

The Bull Case, Tested

The bull case rests on three distinct mechanisms, each testable against empirical evidence rather than assertion.

First is platformization. Changchuan is one of very few global vendors — and effectively the sole domestic Chinese supplier — offering automated testers, mechanical handlers, wafer probers, and optical inspection systems as an integrated test cell. Commercially, an assembly-and-test customer purchasing a complete cell secures a single point of accountability for throughput and yield, avoiding disputes between multiple equipment vendors when line performance falls short.

The acquisition record demonstrates that this strategy is operational rather than hypothetical: STI's machine vision capabilities and EXIS's turret handlers were integrated into Changchuan's equipment lines, followed by the acquisition of Kewei Vision to deepen the software layer.¹⁴⁷¹⁵ Gross margin comparisons, however, highlight the cost of this breadth: Beijing AccoTest operates at a gross margin of roughly 74% on a specialized platform, whereas Changchuan's blended margin sits in the mid-fifties.⁸

Second is the localization mandate, a driver prone to over-extrapolation. While domestic replacement offers a compelling tailwind — given that domestic market share remains below 10% across key digital test categories — industrial policy creates market opportunities rather than guaranteed revenues.³ Policy support also funds domestic competitors, granting Changchuan the chance to contest unserved market share without guaranteeing commercial wins or pricing power.

Third is advanced packaging, the most technically grounded growth driver. Growth in this segment does not require displacing incumbents in legacy architectures, but rather participating in market expansion. Complex multi-die integration increases required test insertions and inspection steps, with high-bandwidth memory structures reportedly requiring over fifteen test passes, compared with three or four for conventional devices.²¹ This structural shift expands addressable demand, particularly for wafer-level probing and optical inspection where Changchuan's recent investments are concentrated. Downstream expansion reinforces this trend, with industry research noting domestic memory expansions and new advanced packaging facilities where test equipment comprised over half of total equipment procurement.²¹

A comparison of domestic business models underscores contrasting strategic choices. AccoTest maintains high unit profitability through segment focus, generating gross margins near 75% with moderate R&D intensity relative to sales.⁸ Changchuan has expanded across four product categories, three continents, and a significantly larger research organization, accepting lower blended margins to build a complete test-cell offering. If chipmakers increasingly prioritize single-vendor test cells and advanced packaging integrates testing and inspection workflows, Changchuan's platform bet will pay off. Conversely, if segment specialization continues to dictate customer purchasing decisions, focused peers will retain superior economics.

The Bear Case

The bear case presents an equally coherent counter-narrative centered on execution, dilution, and financial structure.

First, sustained R&D intensity exceeding 20% of sales may fail to secure high-end market share. Under this scenario, Changchuan maintains the cost structure of an advanced developer while remaining constrained to mid-range equipment segments, competing primarily on price against entrenched incumbents. Recent financial results do not eliminate this risk; a 55% surge in tester revenue remains consistent with expanding mid-range volume rather than breaking into high-performance digital segments.⁴

Second, shareholder dilution remains an ongoing drag on per-share returns. Potential dilution from proposed equity placements — which could lower the founding couple's equity interest from 28.48% to roughly 18.10% — reflects a broader pattern of reliance on capital markets.²³ By funding research and acquisitions through equity issuances, top-line expansion yields diluted per-share earnings growth. Persistent insider share sales alongside capital calls accentuate these governance concerns.⁶¹³

Third, cyclical exposure amplifies downside risk during industry downturns. Semiconductor capital equipment remains highly cyclical, and Changchuan's elevated fixed-cost structure magnifies revenue volatility. Valuation metrics, with forward multiples cited between 44 and 79 times 2026 earnings, reflect expectations of sustained high growth.⁴²¹ Independent analysis indicates these valuations imply approximately 50% annualized profit growth over multiple years, leaving little margin of safety if market conditions soften.²²

The Spine

Changchuan's investment trajectory depends on three conditions: digital and memory testers must expand into a transparent, high-margin revenue stream; the 12-inch wafer prober franchise must successfully transition from image sensors and logic into commercial memory applications; and operating cash flow must normalize to fund expansion without further equity dilution. The thesis breaks if trade controls restrict core sub-assembly sourcing, high-end digital qualifications stall, or an industry downturn materializes before operational cash flow stabilizes.

↑ Back to Top

XII. Key Performance Indicators (KPIs) to Track (03:40:00 – 03:43:00)

Three metrics separate signal from noise in Changchuan's financial trajectory. Everything else is derivative.

First, the high-end tester revenue mix and its associated gross margin. The core investment thesis turns on a structural product shift from handlers and analog testers toward digital SoC and memory testers. Key indicators include the tester segment's share of total revenue, its growth rate relative to the corporate average, and the trajectory of tester gross margins — where the 62.25% recorded in 2025 reflected a 4.36 percentage-point decline.⁴ Whether that margin recovers or continues to erode distinguishes a firm winning on proprietary technology from one competing on price. Disclosing digital tester revenue separately would provide even greater clarity. Independent research has established a specific benchmark: if SoC and memory testers reach over 40% of total revenue, blended gross margin should exceed 60%.²² This threshold offers a concrete test of the bull case. Anchoring on gross margin rather than top-line growth is essential because volume can be purchased through price concessions — particularly in policy-driven markets — whereas stable margins demonstrate that customers lack equivalent alternatives.

Second, 12-inch wafer probe station adoption. This product line represents Changchuan's direct challenge to entrenched international leaders. Progress can be tracked through specific milestones: shipment volumes for the CP12 series, customer qualification counts at major domestic foundries, and whether the CP12-Memory model transitions from customer certification and small-batch promotion into commercial volume production.¹⁸¹⁹ Wafer probers serve as the clearest gauge of technological capability because sub-micron mechanical precision offers no room for compromise: the equipment either maintains alignment and throughput within tight tolerances or fails to qualify. The memory variant remains the critical milestone to monitor, as it represents the firm's most complex technical hurdle and its largest unpenetrated domestic market.

Third, cash conversion efficiency. The ratio of operating cash flow to net profit validates or refutes reported earnings quality. This ratio stood at roughly 0.42 in 2025 and dropped into negative territory in the first quarter of 2026 as the production ramp absorbed working capital.²²⁴ While temporary working capital expansion is standard during steep manufacturing ramps, a persistent deficit indicates structural risk. A rebound toward a range of 0.7 to 1.0 would confirm that reported net profit growth is genuine and self-sustaining. Conversely, continued cash flow weakness paired with rising receivables and inventories would signal that top-line growth relies on balance-sheet expansion rather than operational demand. Ultimately, cash generation will determine whether future equity raises remain strategic choices or balance-sheet necessities.

↑ Back to Top

XIII. Epilogue (03:43:00 – 03:45:00)

There is a version of this story that gets told as a national triumph — the plucky Hangzhou engineering firm that broke a foreign duopoly through sheer persistence and state support. It is a satisfying narrative, but it is premature. What the empirical record actually shows is both more nuanced and more instructive.

Changchuan built its position through a specific, repeatable sequence. It entered through the least defended market segment — mechanical handlers, where the primary hurdle was precision engineering rather than proprietary physics. It used that beachhead to earn something more valuable than short-term gross margin: physical presence on the factory floor, vendor registration, and deep customer intimacy.

The company converted that operational access into qualification in adjacent, higher-barrier categories. It acquired specialized capabilities it could not quickly build internally — optical algorithms in Singapore, turret mechanics in Malaysia, and vision software domestically — paying for them primarily with its own equity. Crucially, management sustained research investments through an industry downturn that would have justified cost-cutting, establishing the foundation for its subsequent operational expansion.

Yet that platform strategy carried substantial costs often omitted from simple growth narratives. Every stage of expansion was funded partly through equity — shares issued to acquire foreign assets, shares sold to fund R&D programs, and equity liquidated by founding leadership. Shareholders who held the stock through the cycle funded technological development while absorbing material dilution. That dynamic reflects the raw arithmetic of attempting to challenge two entrenched global incumbents from a standing start: technology acquisitions and R&D scale require ongoing capital calls.

The unresolved strategic question is whether the final step up the equipment ladder resembles the previous rungs. Transitioning from mechanical handlers to analog testers was an incremental step up a manageable slope. Moving from analog testing into leading-edge digital and memory test platforms represents a climb up a vertical wall, directly against two global leaders whose market share in advanced silicon has expanded during the artificial intelligence cycle. While the domestic addressable market remains vast and largely unpenetrated, the technical gap is real and measurable.

For observers of capital equipment industries, the central lesson centers on sequencing. Hardware platform moats are rarely designed in advance; they are accreted step by step, one qualified customer and one adjacent tool at a time, with cash-generative entry products funding higher-risk R&D next door. Changchuan's mechanical handlers funded its analog testers. Its tester business is now funding its wafer prober expansion. Whether its digital platforms and probe stations ultimately deliver self-sustaining returns will be decided not by policy directives, but by gross margins, qualification milestones, and cash flow conversion.

References

1. 从士兰微到长川科技，12位亿万富豪的终极PK — Sohu, 2017-09-19 ↩↩↩↩↩

2. 存储行业、测试设备、长川科技 — Sina Finance, 2026-06-25 ↩↩

3. 2025年中国半导体测试机行业市场现状分析：高端领域仍以海外厂商为主，国产替代空间清晰 — 华经产业研究院 Huaon, 2025 ↩↩↩↩↩↩↩

4. 长川科技2025年报&2026一季报点评：业绩持续高增，充分受益国产替代加速&封测厂扩产 — Sina Finance, 2026-04-27 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

5. 长川科技2025年净利润大增190% 2026年一季度再增217% — Securities Times (证券时报), 2026-04-24 ↩↩↩↩↩↩

6. 核心技术又高又硬，内部人套现30亿又快又准！长川科技：国产替代最前端，技术与减持齐头并进 — Sohu, 2024 ↩↩↩↩↩↩↩↩↩↩

7. 业绩短暂波动后快速恢复正常，长川科技制胜秘诀有哪些？ — Sina Finance, 2025-01-06 ↩↩↩↩↩

8. 华峰测控：2025年净利润同比增长60.55% 拟10转4.8股派11.9元 — Eastmoney, 2026-04-29 ↩↩↩

9. Cninfo Official Financial Disclosure Portal for Changchuan Tech (300604) — Cninfo ↩

10. 长川科技一箭双雕：大基金跃升第二大股东、拿下长新投资90%股权 — DRAMeXchange (全球半导体观察), 2018-12-13 ↩↩↩

11. 长川科技2024年营业收入升105.2%，净利润升915.1% — Sina Finance, 2025-04-29 ↩↩↩↩

12. Hangzhou Changchuan Technology 2025 Annual Report — Hangzhou Changchuan Technology, 2026-04-27 ↩↩↩↩

13. 长川科技股东杭州长川投资管理合伙企业（有限合伙）拟减持1196.84万股套现9.23亿元 — Sina Finance, 2025-11-28 ↩↩↩

14. 长川科技之子公司STI是什么来头？ — Eastmoney, 2025-01-18 ↩↩↩↩

15. 长川科技"两步走"2.16亿全控科为升 提升技术实力研发投入占营收26.65% — Sina Finance, 2025-09-10 ↩↩

16. Advantest Corporation Investor Relations — Advantest ↩

17. Teradyne Inc. Investor Relations & Financial Reports — Teradyne ↩

18. 长川科技：在探针台研发及产业化项目实施过程中，当前CP12-SOC/CIS/Discrete等项目已达到量产状态 — 同花顺财经 10jqka, 2025-11-24 ↩↩↩↩↩

19. 长川科技：部分探针台产品已达量产状态，CP12-Memory项目处于客户端认证和小批量推广阶段 — Sohu, 2025 ↩↩↩↩↩↩

20. 长川科技：预计2026年1-6月归属净利润盈利9亿元至10亿元 — Sohu, 2026-06-22 ↩↩↩↩

21. 长川科技2026半年报业绩预告点评：Q2业绩超市场预期，国产测试机龙头SoC&存储双轮驱动 — Sina Finance, 2026-06-23 ↩↩↩↩↩

22. 长川科技深度：拆解2026Q1的"高增长质量"，以及未来两年的三个关键变量 — Eastmoney, 2026-04-24 ↩↩↩↩↩↩↩↩↩

23. 长川科技发布定增预案 实控人持股或被稀释至约18% — Securities Times (证券时报), 2025-06-24 ↩↩↩

24. 长川科技2025年度定增获深交所审核通过 拟募资不超31.27亿元加码半导体设备研发 — Sina Finance, 2026-03-11 ↩↩↩