Piotech Inc: The Story of China's Thin-Film Deposition Champion

I. Introduction & The Core Thesis

Every advanced chip in the world is, in a literal sense, a sandwich. Beneath the transistors and above them sit dozens — in the case of 3D NAND memory, hundreds — of films so thin they are measured in atoms. Some insulate. Some conduct. Some exist only to stop one layer from poisoning the next. None of them can be poured on. They have to be grown, molecule by molecule, inside a vacuum chamber, at a temperature low enough that the circuitry already etched underneath does not cook.

The machines that do this growing are called deposition tools, and for most of the industry's history they were made in Santa Clara and Fremont, California. Applied Materials, Lam Research, and Tokyo Electron together account for roughly seven-tenths of the global chemical vapour deposition market — Applied at around 30%, Lam at 21%, Tokyo Electron at 19% by one widely cited breakdown from the period around Piotech's listing.¹ Thin-film deposition is one of the three great cost pools of wafer fab equipment alongside lithography and etch, and unlike lithography — where ASML's monopoly is absolute and Chinese substitution remains a distant project — deposition was always the pillar where a domestic challenger might actually get somewhere.

Deposition accounts for roughly a fifth of what a modern fab spends on tools. But the more important number is the count. A fab buys a handful of lithography scanners and hundreds of deposition chambers, spread across dozens of distinct tool types and process variants. That fragmentation is precisely what makes the category attackable: a challenger does not need to build the best deposition tool in the world; it needs to build one that is good enough for a specific film, on a specific layer, at a specific node — and then repeat that exercise a hundred times.

That challenger is 拓荆科技 Piotech Inc, listed as 688072.SS on the 上海证券交易所 Shanghai Stock Exchange's 科创板 STAR Market. It was founded in April 2010 in Shenyang — an old northeastern industrial city better known for machine tools and heavy engineering than for semiconductors — by a group of Chinese-American engineers who had spent their careers at Intel, Novellus Systems, and other Western equipment houses, in partnership with an enterprise affiliated with the 中国科学院 Chinese Academy of Sciences.²

Sixteen years later the numbers are large. For the 2025 financial year, Piotech reported revenue of ¥6.519 billion, up 58.87% year over year, and net profit attributable to shareholders of ¥927 million, up 34.67%.³ Order backlog stood at roughly ¥11 billion — about 1.7 times a full year of revenue — with contract liabilities, essentially customer prepayments, of ¥4.852 billion, themselves up 62.66%.³ Cumulative shipments passed 3,400 reaction chambers into around 100 integrated-circuit production lines.³ In the first half of 2026, revenue grew another 49.06% to ¥2.913 billion.⁴

Here is where a careful reader should slow down. That 2025 profit grew far more slowly than revenue, and the reason is straightforward: gross margin fell 6.90 percentage points to 34.01%, because cost of goods sold grew 77.21% against revenue growth of 58.87%.⁵ Two years earlier, gross margin had been above 50%.⁶

In plain terms, a company supposedly protected by ferocious switching costs and proprietary process know-how watched roughly seventeen points of gross margin evaporate over two years — during the exact period when volumes were exploding and scale economies were meant to be arriving.

The headline profit figures require the same scrutiny. The 1,324% net profit growth reported for the first half of 2026 was overwhelmingly an accounting event: ¥954 million of the ¥1.343 billion headline total came from non-recurring items, principally gains on securities Piotech holds in its own supply chain, of which ¥902 million arose from the STAR Market flotation of a single RF power-supply maker it had backed years earlier.⁴ Strip that out, and first-half operating profitability was ¥367 million.⁴

So the story that follows is not a simple narrative about a national champion riding an unstoppable policy wave. It is a story about a company that has genuinely done something very hard — building qualified, production-worthy deposition tools that run on live 12-inch fab lines in China — while operating under three permanent constraints: a customer base so concentrated that the top five accounted for 79.15% of 2025 revenue,³ a business model that ships tools to customers and then waits a year or more to be paid, and a component supply chain that runs straight through the jurisdictions that put Piotech on the US Entity List in December 2024.⁷

The revenue is dominated by one product family. PECVD — plasma-enhanced chemical vapour deposition — generated ¥5.142 billion in 2025, up 75.27%, or roughly 79% of the total.³ ALD atomic layer deposition, the tool line that matters for the most advanced nodes, generated ¥301 million, up 191.82% — real growth from a small base, accounting for about 4.6% of revenue.³ Hybrid bonding equipment, built by the subsidiary 拓荆键科 Piotech Keytech and aimed at the stacked-chip architectures behind high-bandwidth memory, generated ¥136 million, up 41.92%, or about 2.1%.³ The remainder — HDPCVD, SACVD, Flowable CVD, and related gap-fill specialties — makes up the balance.

That composition matters. Piotech is, today, a one-product company with two interesting options attached. The core investment question is whether the main product's economics are durable, and whether those options convert. The company's own history offers hard evidence on both.

Myth versus reality. Three consensus beliefs about Piotech are worth checking before the story begins.

Myth: it out-engineered the Americans. Reality: Piotech does not compete at the leading edge of global deposition technology, does not sell meaningfully outside China, and does not publish a tool-for-tool performance comparison against the incumbents it displaces. What it did — genuinely difficult and rare — was reconstruct a class of production-grade equipment inside a market that had every incentive to buy it, and survive the ten unprofitable years that reconstruction required.

Myth: sanctions were an unambiguous windfall. Reality: the same policy regime that handed Piotech its customers also placed the company and four of its subsidiaries on the US Entity List, constraining the components it buys rather than the tools it sells.

Myth: state ownership is a permanent floor. Reality: the state fund that anchored the company through its loss-making decade has been reducing its stake into strength, while a successor fund buys the next generation of assets. The register is a portfolio, not a pledge.

The essential analytical questions all sit downstream of those corrections: what such a business earns, who captures the surplus, and how much of the current valuation rests on outcomes that have not yet materialized.

One framing device is useful throughout. In wafer fab equipment, there are really only two questions an owner ever needs answered: how many chambers does this company have running in production, and what does it earn on each one? Piotech has spent sixteen years answering the first question emphatically. The second remains open, and 2025 was the year the market first received an uncomfortable partial answer.

The narrative spine, then, runs like this: a group of returnee engineers built something in a city with no semiconductor industry. The state paid for the decade in which it earned nothing. Geopolitics handed it a customer base. And now, at scale, the business is being tested on the only question that ultimately matters for an owner: whether a hard-won technical position converts into durable economics, or merely into volume at whatever price four buyers decide to pay.

↑ Back to Top

II. Founding Context & The Returnee Pioneers (2010–2014)

Picture Shenyang in 2010: the capital of Liaoning, the heartland of China's heavy-industry rust belt, defined by state-owned machine tool plants and harsh winters. It sits roughly 2,000 kilometres north of Shanghai's developing semiconductor hub and far removed from Silicon Valley. Yet in April 2010, a team of engineers who had spent decades inside the American semiconductor equipment industry arrived in Shenyang to build a wafer fab equipment maker from scratch.²

The senior technical lead was 姜谦 Jiang Qian, born in 1952, a US citizen with a doctorate from Brandeis University who worked in semiconductor process development at Intel and Novellus Systems before joining Piotech in 2010.⁸ Novellus played a central role in this lineage: it developed the modern high-productivity PECVD platform, using a sequential multi-station architecture that allowed one chamber to process several wafers at once, before being acquired by Lam Research in 2012. Jiang brought direct experience in the design philosophy that Piotech would spend the next decade reconstructing from first principles.

A process engineer in Silicon Valley or Oregon in 2010 operated at the core of the global ecosystem, supported by a mature supply chain, a dense talent pool, nearby customers, and a consolidating equipment sector positioned for high margins.

Moving to Shenyang meant leaving that foundation behind. It required hiring engineers who had never seen a production PECVD chamber and sourcing precision components from domestic suppliers that were, at the time, virtually non-existent.

The move rested on a long-term calculation. The founders anticipated that China would commit substantial resources to domestic semiconductor manufacturing over the subsequent two decades, and that equipment—the most technically complex layer with the highest barriers to entry—would represent both the final frontier of localization and the most valuable asset if established successfully. They were positioning for a multi-decade policy trajectory rather than a single product cycle.

In 2014, 吕光泉 Lv Guangquan joined the company. Born in 1965, Lv held a doctorate and had served in senior roles at Novellus in the United States and at an equipment manufacturer in Germany before joining Piotech as technical director, later becoming general manager, director, and eventually chairman.⁸ The partnership between a returnee technical founder and a returnee technical operator reflected the prevailing structure among early Chinese semiconductor equipment startups.

Why deposition, and why plasma? Semiconductor manufacturing involves sequentially depositing layers of material onto a silicon wafer and etching patterns into them. Depositing insulating or conductive films such as silicon dioxide or silicon nitride is straightforward at high temperatures, where process gases thermalize and decompose on the wafer surface.

However, as a chip reaches upper layers, high heat risks damaging the underlying transistors. Plasma-enhanced chemical vapour deposition solves this by using radio-frequency energy to break process gases into reactive fragments, allowing chemical reactions to occur on the wafer surface at a few hundred degrees Celsius instead of a thousand. This lower-temperature process enables multi-layer modern chip manufacturing.

Consequently, PECVD tools are among the most heavily utilized machines in a fab, with advanced logic and 3D NAND lines passing wafers through deposition steps dozens of times. While lithography scanners are few in number and capital-intensive per unit, deposition tools are deployed in large quantities across a fab line. A domestic supplier that qualifies a subset of those process slots can capture significant initial revenue. However, tool reliability must be exceptionally high, as particle contamination in a single chamber at a late processing stage can invalidate the entire wafer.

From a development standpoint, deposition tools differ fundamentally from microprocessors. They are integrated electromechanical systems combining vacuum chambers, gas delivery manifolds, RF plasma sources, robotic wafer handlers, and control software. The underlying physical principles are established in public literature and expired patents.

The primary engineering challenge lies in integration and process control: maintaining uniform film deposition across thousands of wafers while managing mechanical and environmental drift over years of operation. Solving this requires patient engineering and extensive run-rate data on active production lines. This iterative profile made deposition a more accessible entry point for a domestic challenger than advanced lithography.

Piotech's initial capital structure reflected the extended development timeline required for semiconductor capital equipment. The company was founded as a joint venture between the returning engineering team and an enterprise affiliated with the Chinese Academy of Sciences, and it received national high-tech enterprise status early in its operations.² Its development of a 12-inch PECVD tool was supported through national science and technology programmes, with the company participating in nine such projects by mid-2024.⁹

Chairman Lv noted that through these national projects, the company "mastered core technologies for PECVD thin film deposition equipment and accumulated experience in product industrialisation."⁹ The initial emphasis remained on technical validation and process experience rather than immediate commercial scale or customer adoption.

This structure established a long-term governance arrangement. Piotech was organized not as a traditional venture-backed startup, but as a joint venture pairing technical management with institutional capital. Management retained technical oversight, while state-affiliated institutions provided capital and board representation, creating an operational structure that required alignment between institutional stakeholders and executive leadership.

Testing the founding myth: did technical achievement build a business? The traditional narrative suggests that returnee technical expertise combined with state R&D backing immediately yielded a commercially viable equipment supplier. The financial record indicates a longer verification period.

In 2018, eight years after its founding, Piotech generated ¥70.64 million in revenue—roughly $10 million, or less than the cost of a single advanced lithography scanner—and recorded a net loss of ¥103 million.¹⁰ In 2019, revenue rose to ¥251 million while net losses narrowed to ¥19.37 million; in 2020, revenue reached ¥436 million alongside a net loss of ¥11.70 million.¹⁰ The company's IPO prospectus attributed these extended losses to high research intensity and multi-year customer verification cycles.¹¹ Independent analysis of the filings showed R&D expenditure consuming between 28% and 153% of annual revenue from 2018 through 2021.¹

This timeline highlights the structural realities of semiconductor capital equipment. Piotech required a decade from inception to generate meaningful commercial revenue, funded primarily by state-linked equity and development grants rather than commercial tool sales. This historical baseline provides context for evaluating Piotech's expanding product lines, such as atomic layer deposition and hybrid bonding, where the transition from technological validation to commercial scale requires multi-year lead times.

The eventual acceleration in commercial adoption stemmed less from a single laboratory breakthrough than from shifts in domestic customer purchasing requirements.

↑ Back to Top

III. The 02 Special Project & The Big Fund Catalyst (2014–2019)

There is an unglamorous moment in the life of every semiconductor equipment maker that determines whether it survives, and it is not a product launch. It is the day a fab engineer agrees to give up a slot on an active production line — a slot currently occupied by a working tool from Applied Materials — to let an unproven machine run wafers carrying live customer product.

For a Chinese deposition challenger in the mid-2010s, securing that slot required two assets Piotech could not manufacture in Shenyang: substantial capital and trusted industry introductions.

The difficulty of securing an evaluation slot is hard to overstate. A 12-inch wafer fab represents billions of dollars in capital expenditure and operates continuously. Its economics depend heavily on yield — the percentage of functional die per wafer — which can be ruined by microscopic particles, film thickness variations, or trace contamination.

A deposition tool that introduces just one extra particle per wafer can cost a fab more in scrapped product over a single month than the purchase price of the machine itself. Consequently, process engineers are risk-averse by training. When presented with an unproven tool from an unverified vendor, their default response is to decline — and for most of the 2010s, that refusal was economically rational.

The necessary funding came through China's state R&D architecture. The 国家科技重大专项 national major science and technology programmes for integrated circuit manufacturing equipment — known colloquially as the 02专项 02 Special Project — were structured specifically to bridge the valley of death between a working prototype and a fab-qualified tool, a phase characterized by heavy expenditure and zero revenue.⁹ Rather than a conventional commercial subsidy, this funding represented state support for the option value of establishing a domestic thin-film deposition capability.

The required introductions came through state-backed equity. The 国家集成电路产业投资基金 China Integrated Circuit Industry Investment Fund — commonly known as the Big Fund — acquired a stake that made it Piotech's dominant shareholder, holding 19.86% at the time of the company's IPO.¹ Sitting alongside it was an unusual strategic investor: 中微公司 AMEC, the Shanghai-based etch equipment manufacturer founded by 尹志尧 Gerald Yin, which held an 8.4% stake.¹ Yin himself took a seat on Piotech's board.⁸

This capital structure lacked a direct Western equivalent. The presence of a major etch vendor holding a significant stake and a board seat in a deposition startup reflected how China assembled its equipment sector: not as competing entities fighting for the same accounts, but as a coordinated ecosystem of complementary process capabilities anchored by state capital, with direct corporate rivalry deferred until participants reached scale.

Backed by the Big Fund, Piotech gained entry to domestic production lines. Its initial customer base concentrated around the primary state-backed wafer manufacturers: 中芯国际 SMIC, 华虹集团 Hua Hong Group, 长江存储 YMTC, and the DRAM producer 长鑫存储 CXMT.¹

This qualification push coincided with a favorable expansion cycle. In the late 2010s, China's memory strategy transitioned from planning to construction, with greenfield facilities breaking ground at YMTC's 3D NAND complex in Wuhan and CXMT's DRAM fab in Hefei. Greenfield lines present far lower entry barriers for new equipment suppliers than mature fabs, as there are no legacy tools to displace and no established production yields to jeopardize.

Furthermore, 3D NAND manufacturing is exceptionally deposition-intensive. Building memory cells vertically requires depositing and etching dozens or hundreds of alternating thin-film layers, resulting in significantly higher deposition tool density than in logic fabs. Piotech's core PECVD product aligned directly with the fastest-growing segment of domestic tool demand.

What "qualification" actually costs. The tool qualification process demonstrates where competitive moat theory meets operational cash flow. When Piotech wins an evaluation slot, it does not record an immediate sale; it ships a machine.

The tool is delivered to the customer's fab, installed, calibrated, and evaluated over several months. Fabs measure particle counts, film thickness uniformity across the wafer, film stress, mechanical throughput, and mean time between failures. Each metric has strict operational thresholds; failing any test requires engineering modifications. Piotech recognizes revenue only after the customer formally signs off on tool acceptance.¹² The company describes a commercial cycle spanning initial tender, delivery, installation, debugging, customer acceptance, and subsequent after-sales support — a structure where cash realization occurs near the end of the process.¹²

The financial impact of this extended cycle appears in the company's filings under 发出商品 — goods shipped but not yet accepted. The balance stood at ¥1.291 billion at the end of 2022, rose to ¥1.902 billion by year-end 2023, reached ¥4.134 billion at the close of 2024, and hit ¥4.649 billion by mid-2025.¹² By the end of 2025, total inventory reached ¥7.826 billion, composed of ¥4.362 billion in shipped-but-unaccepted tools and ¥621 million in work in progress.¹² At year-end 2025, Piotech held uncollected tools at customer sites equivalent to roughly two-thirds of its full-year revenue.

Piotech's disclosures outline the risks inherent in this working-capital structure: if tool acceptance encounters delays or rapid technology shifts occur, the realizable value of inventory could fall below book value, necessitating impairment charges against net profit.¹²

Testing the claim that qualification wins are capital-efficient growth. Bullish narrative framings treat the shipped-goods balance as deferred revenue — a backlog of tools awaiting conversion into cash. However, financial data shows that shipped inventory grew faster than recognized revenue for four consecutive years, indicating that the conversion queue expanded rather than cleared. This working-capital burden is not a temporary entry cost from the initial qualification push; it represents a permanent structural feature that scales alongside revenue growth.

A rigorous assessment does not dismiss the strategic value of Piotech's installed base, which underpins its competitive position. However, the business model converts shipments into cash with a multi-quarter lag. A sharp drop in customer demand or delays in accepting next-generation tool lines would register on the income statement as inventory write-downs rather than minor revenue slowdowns. Key operational metrics to monitor are whether shipped inventory begins growing slower than recognized revenue over sustained periods and whether material inventory impairments occur.

Additionally, revenue timing remains partly at customer discretion due to acceptance-based recognition rules, while shipped inventory valuation relies on management estimates of acceptance probability. While public filings show no evidence of improper accounting practices, these line items reflect significant management judgments that materially impact reported profitability.

By 2019, the commercial environment for domestic fab procurement shifted fundamentally — driven less by incremental engineering progress than by policy moves in Washington.

↑ Back to Top

IV. Escalating Sanctions & The Domestic Self-Reliance Inflection (2019–2022)

For most of the 2010s, a Chinese fab purchasing a domestic deposition tool was making an operational compromise justified primarily by cost. The domestic tool was less proven and less expensive, relegated to non-critical layers where it served mainly to give procurement teams leverage in price negotiations with Western incumbents like Applied Materials. That strategy endured as long as foreign equipment remained readily available.

Between 2019 and 2022, that availability ended. As US export restrictions targeted Huawei's chip design unit, SMIC, and later YMTC, domestic fab operators were forced to restructure their expansion plans around the risk that critical tools, spare parts, or field services could be cut off. Achieving a self-controllable supply chain (自主可控) shifted from a policy aspiration into an urgent operational requirement in fab capital planning.

That shift altered the industry's vendor base—the multi-year commitment fabs make when qualifying a tool supplier for a specific process layer. Once a vendor is qualified, process integration teams build the surrounding manufacturing flow around that tool's specific physical characteristics. Consequently, fabs alter their vendor base rarely and deliberately, usually only under strong external pressure.

This realignment proved to be the pivotal inflection point for Piotech, outstripping any individual product launch in significance. Fabs that had previously restricted domestic tools to non-critical layers—such as nitride passivation films or low-stress liners, where a failure ruined a single wafer rather than a quarter's output—began evaluating them for primary slots on 3D NAND stacks and 28-nanometer and 14-nanometer logic lines. Winning primary slots offered a dramatic increase in revenue per fab, but it also imposed stringent technical demands: tools had to deliver high yield in continuous production, not just in short demonstration runs.

The financial impact was immediate. Revenue rose sharply from ¥436 million in 2020, positioning Piotech for a public listing. On April 20, 2022, the company listed on the STAR Market, issuing 31.619 million shares at ¥71.88 per share. The offering raised ¥2.273 billion in gross proceeds and ¥2.128 billion in net proceeds—yielding approximately ¥1.127 billion more than its original target.¹³ The strong over-subscription reflected an eager market ready to fund domestic semiconductor equipment capabilities at high valuations.

The vulnerability that the sanctions narrative obscures. While popular narratives frame US export controls purely as a tailwind for domestic suppliers, the empirical record presents a more complex reality: Piotech is both a beneficiary and a target of trade sanctions.

On December 2, 2024, the US Bureau of Industry and Security added 136 Chinese entities to the Entity List in a targeted action against semiconductor manufacturing equipment suppliers.⁷ Piotech was listed alongside major domestic peers, including 北方华创 Naura, 盛美上海 ACM Research Shanghai, 中科飞测 Skyverse, 华海清科 Hwatsing, and 芯源微 Kingsemi.⁷¹⁴ The designation applied to the parent company and several key subsidiaries, including its Beijing, Shanghai, and Haining units. Management stated publicly that the listing would not disrupt day-to-day operations, pointing to its stable supply chain.¹⁵

That reassurance requires careful analysis. Stating that operations face no immediate disruption addresses current inventory levels, but it does not address the longer-term structural mechanism.

A deposition tool functions as a complex integrated system. Its most critical components are not the chamber shell, but the radio-frequency power supplies that control the plasma field, mass flow controllers that regulate process gases with high precision, turbo pumps, vacuum seals, and control software. Historically, leading suppliers of these components have been based in the US, Japan, and Europe. While the Entity List does not prohibit Piotech from selling tools to domestic customers, it restricts the components it can import. That limitation presents a persistent challenge that standard twelve-to-eighteen-month component buffers cannot permanently resolve.

To address this vulnerability, Piotech invested directly in its domestic supply chain. The company became an anchor customer and equity investor in 恒运昌 Hengyunchang, a Shenzhen-based producer of plasma RF power supply systems—components described in domestic industry literature as the precise tool controlling the plasma field.¹⁶

By 2024, Hengyunchang secured the leading domestic market share for semiconductor plasma RF power supplies in China, expanding its revenue from ¥158 million in 2022 to ¥541 million in 2024. The supplier completed an IPO on the STAR Market on January 28, 2026, as the exchange's first listing of that year.¹⁶ At the time of Hengyunchang's prospectus, Piotech held a 3.42% direct and indirect stake while accounting for 45% to 63% of Hengyunchang's annual revenue during the reporting period.¹⁷

This arrangement illustrates strategic supply-chain management at the firm level: Piotech mitigated component risk by supporting a key domestic supplier, while capturing financial gains when that supplier went public. However, it also creates an interdependent structure where the supplier relies heavily on a single customer, and where Piotech recognized a ¥902 million investment gain in the first half of 2026 from an entity whose revenue it largely generates.⁴ Evaluating the company requires considering both the strategic benefit and the concentration risk simultaneously.

The personnel dimension. In October 2022, US regulatory updates restricted US citizens and permanent residents from supporting certain advanced semiconductor manufacturing activities in China, creating operational challenges for domestic equipment firms founded by returnee engineers. Piotech's subsequent disclosures noted leadership realignments. On October 29, 2024, founder Jiang Qian—then 72—and senior technical director Ye Wumao stepped down from their designated core technical personnel roles. On December 14, 2024, chairman Lv Guangquan and former general manager Tian Xiaoming also departed the core technical register, with filings citing "work adjustment."¹⁸ Jiang remains a director and Lv continues as chairman.⁸¹⁸

The company's filings do not attribute these changes to foreign export controls. Nevertheless, the timing is notable: within a fourteen-month window surrounding the Entity List designation, Piotech's four senior technical leaders transitioned out of formal core-technical-personnel designations while retaining governance or advisory positions. For a firm built on imported process expertise, the seamless transfer of technical leadership remains an essential factor that public disclosures do not fully detail.

Beyond formal R&D metrics lies the transfer of tacit engineering knowledge: the practical judgment required to refine chamber designs, diagnose yield variations on active fab lines, and resolve process conflicts with customers. While Piotech has expanded its engineering workforce rapidly, scaling headcount does not automatically replicate decades of specialized industry experience.

Because filings do not track tacit knowledge transfer directly, external observers must monitor indirect indicators—specifically, the commercialization pace of new tool categories and field reliability rates across the expanding installed base.

By 2022, Piotech possessed a public listing, strong policy alignment, and a product portfolio that domestic fabs were actively integrating into production lines. The central question shifted to whether this expanded market position could generate sustainable economic returns.

↑ Back to Top

V. The Core Engine: PECVD Economics, Fab Qualification & Market Share

On a 12-inch wafer fab line, plasma-enhanced chemical vapour deposition tools are the primary workhorses: grey rectangular cabinets housing multiple process chambers clustered around a central robotic handler, operating around the clock. Their commercial proposition is straightforward to state yet demanding to execute. A semiconductor manufacturer purchases a tool once and relies on it for a decade, meaning the transaction is ultimately an investment in uptime, yield, and process repeatability—ensuring that the film deposited in a given chamber today matches the film produced months earlier.

Piotech's core PECVD division has reached significant scale. The ¥5.142 billion in revenue generated by the line in 2025 represented a 75.27% year-over-year increase—outpacing overall company growth and expanding its share of the product mix even as newer equipment categories expanded from smaller bases.³ By the end of 2025, cumulative chamber shipments exceeded 3,400 across approximately 100 production lines, expanding past 3,800 chambers and 457 million cumulative wafer passes by mid-2026.³⁴ This installed base constitutes the company's most vital asset.

Piotech measures its volume in process chambers rather than complete tool frames. In deposition architecture, a platform uses a central vacuum robot to service multiple surrounding process chambers. Increasing the number of chambers per platform allows a toolmaker to maximize wafer throughput per square metre of cleanroom floor—a critical efficiency metric given the high capital cost of cleanroom real estate. Chamber counts serve as a direct measure of revenue, productivity, and the depth of vendor integration within customer production lines.

Where the moat actually lives. A deposition process recipe is far more complex than a standard machine setting. It represents an intricate calibration of gas flow rates, radio-frequency power, chamber pressure, wafer temperature gradients, and chamber-conditioning cycles, refined over thousands of wafer runs to achieve specific film properties for a customer's target process node. Because this process map is co-developed alongside the fab and tied directly to the tool's physical chamber geometry, it cannot be transferred to alternative hardware. Replacing an established PECVD tool requires repeating the entire qualification process—consuming months of engineering capacity, reducing line availability, and introducing yield risks on high-value production wafers.

This structural dynamic creates a substantial switching barrier, evidenced by the 457 million cumulative wafer passes recorded on Piotech tools. Each processed wafer adds operational data to the company's process library, building an empirical foundation that new market entrants cannot easily replicate.

However, Western semiconductor equipment business models rely heavily on an additional economic engine that Piotech does not break out separately: the aftermarket. Established manufacturers such as Applied Materials and Lam Research derive stable, high-margin revenue from spare parts, chamber refurbishments, and long-term service agreements across their installed bases. This recurring revenue stream provides a counter-cyclical buffer, as fabs continue servicing existing equipment during capital spending downturns. Although Piotech's installed base has reached a scale where a comparable service annuity should be emerging, the company does not disclose aftermarket metrics separately, leaving its contribution to margin stability unverified.

But switching costs are not pricing power, and Piotech's own margin history proves it. This structural limitation represents a critical analytical insight: high switching costs do not automatically grant pricing power.

The trajectory of Piotech's gross margin illustrates this tension. The company recorded a gross margin above 50% in 2023,⁶ which declined to approximately 41% in 2024.⁶ In 2025, gross margin fell further by 6.90 percentage points to 34.01%, as cost of goods sold expanded by 77.21% against revenue growth of 58.87%.⁵ Despite rapid volume expansion—a period when scale economies and an established installed base were expected to expand margins—the company experienced a seventeen-percentage-point gross margin compression over two years.

During the annual results briefing on July 8, 2026, executive management attributed the 2025 margin compression to an unusually high concentration of new product introductions and process validations, which incurred elevated initial verification costs; Chairman Lv stated that advanced tool lines had entered volume production, pointing to a gross margin rebound in the first quarter of 2026 and projecting subsequent margin stabilization.¹⁹ Management explicitly rejected assertions of price discounting to capture market share, maintaining that tools compete on performance and operational efficiency while citing tool uptime exceeding 90%.¹⁹

The recovery in first-quarter 2026 gross margin to 41.69% and a nine-percentage-point year-over-year margin improvement in the first half of 2026 lend support to management's explanation.⁶⁴ However, two structural factors qualify this rebound. First, attributing margin compression to verification expenses for new tools implies that customer qualification costs recur systematically with every technology transition rather than representing a one-time entry fee. Second, even the recovered 41.69% margin remains well below 2023 levels, and management has not issued guidance indicating a return to prior peaks.

This margin performance must also be evaluated within the broader capital goods cycle. Semiconductor equipment procurement is inherently cyclical, with fab operators adjusting tool purchases in concentrated waves based on long-term capacity plans. Piotech has operated primarily during an extended, policy-supported expansion of domestic fab capacity. Consequently, its historical metrics—including order growth, customer prepayments, tool acceptance rates, and realized pricing—have been established under expansionary market conditions, providing a limited baseline for performance across a complete industry cycle.

On market share, the honest answer is that nobody outside the fabs knows precisely. Industry commentaries frequently attribute a dominant domestic market share to Piotech's PECVD line, with estimates often citing figures above 30%. However, official corporate disclosures do not report specific market share metrics. Verified historical data from the period surrounding the company's IPO indicated more modest positions: a 6.96% share of China's PECVD market, 1.27% in ALD, and 3.25% in SACVD, within a domestic PECVD market then valued at approximately $1.5 billion.¹ Contemporary industry analyses estimated the overall domestic localization rate for thin-film deposition tools at roughly 5.5%, split between Piotech at 3.8% (concentrated in PECVD) and Naura at 1.7% (focused primarily on PVD).

While rapid revenue growth indicates that Piotech's current market share has expanded beyond those early baselines, precise market share figures remain undisclosed by the company. What public disclosures do document is significant customer concentration: Piotech's top five customers generated ¥5.160 billion in 2025, accounting for 79.15% of total annual revenue.³

Testing the moat against buyer power. This high concentration places substantial commercial influence in the hands of a small group of state-backed chipmakers. These procurement teams actively maintain multi-vendor sourcing strategies to prevent single-supplier dependence, balancing domestic tool allocations across competing suppliers.

The domestic competitive landscape reflects this buyer strategy. 北方华创 Naura Technology Group (002371.SZ), which recorded first-half 2025 revenue of ¥16.142 billion across etch, deposition, thermal processing, wet cleaning, and ion implantation, continues to expand its chemical vapour deposition offerings from a broader platform base.²⁰ 中微公司 AMEC (688012.SH)—a strategic equity holder in Piotech—has similarly expanded from etch into CVD and ALD. Meanwhile, 盛美上海 ACM Research Shanghai (688082.SH) has broadened its portfolio from wet cleaning and electroplating into thermal furnace systems.

These domestic competitors present distinct strategic challenges. As a multi-product platform company, Naura operates across etch, PVD, thermal processing, wet clean, and ion implant, enabling it to offer integrated tool suites and bundle deposition products alongside established tool lines—a structural advantage against a specialist supplier focused primarily on thin-film deposition.

AMEC entered deposition after establishing a strong market position in plasma etch equipment; its dual role as a direct product competitor and a strategic shareholder creates a complex commercial relationship. ACM Research Shanghai poses a less immediate overlap, having built its market presence in wet cleaning and plating before expanding into thermal furnace applications.

Against these competitors, Piotech's primary strength lies in its process specialization within thin-film deposition. However, as fab operators seek to streamline vendor management by favoring suppliers with broader equipment portfolios, Piotech faces structural pressure to expand its product range beyond its core deposition lines.

The empirical evidence indicates that while qualified tool slots provide high customer retention across installed platforms, they do not guarantee long-term pricing power. Each new process node or line expansion requires competitive bidding against alternative qualified vendors, resetting qualification requirements and cost structures.

Consequently, evaluating the durability of Piotech's competitive position depends on future margin trajectories. If its process position confers genuine pricing power, gross margins should trend back toward historical highs as newer tool platforms reach mature production volumes and verification costs normalize. Conversely, if gross margins stabilize in the high-30% to low-40% range alongside expanding revenue, Piotech operates as a high-volume hardware supplier within a concentrated, buyer-dominated market.

This dynamic focuses attention on whether the company's emerging equipment categories can deliver higher-margin growth.

↑ Back to Top

VI. Beyond PECVD: The Optionality Stack — ALD, HDPCVD & Hybrid Bonding

In September 2025, state-backed fund 国投集新 Guotou Jixin—an investment vehicle of 国家集成电路产业投资基金三期 Big Fund Phase III—agreed to inject up to ¥450 million into Piotech's hybrid bonding subsidiary, 拓荆键科 Piotech Keytech, at a pre-money valuation of ¥2.5 billion.²¹ For the first half of 2025, Keytech had generated ¥1.1455 million in revenue—not billion, but million—while recording a net loss of ¥37.47 million on net assets of ¥55.22 million.²¹ The valuation represented a premium of roughly 4,427% over book.²¹ It marked Big Fund Phase III's first semiconductor project investment.²¹

The broader financing round raised up to ¥1.03 billion from six investors. Piotech contributed up to ¥450 million and an employee holding platform provided up to ¥100 million. Following the transaction, Piotech's equity stake in Keytech dropped to roughly 53.57%, with Guotou Jixin becoming the second-largest holder at about 12.71%.²¹²²

That transaction provides a clear case study in how Piotech's growth options should—and should not—be valued.

Sub-segment one: atomic layer deposition. If PECVD is akin to spray-painting a wall evenly, atomic layer deposition (ALD) resembles dipping that wall into a solution one molecule thick, rinsing it, and repeating the process thousands of times. Each cycle deposits a single atomic layer. While slow and expensive, ALD produces precise conformality: the film coats the inside of a hole a hundred times deeper than it is wide, uniformly, all the way to the bottom. That property renders ALD essential for DRAM capacitor dielectrics, spacer layers in multi-patterning schemes that allow fabs to print features finer than lithography alone can resolve, and gate stacks at advanced logic nodes. Globally, the benchmark suppliers are ASM International and Lam Research.

Piotech's ALD revenue reached ¥301 million in 2025, up 191.82% year over year, making it the fastest-growing line in the company.³ Yet after more than a decade of corporate history and years of targeted R&D—the company's ALD and SACVD lines registered meaningful sales growth only in 2023, around the time thermal ALD and HDPCVD achieved their first industrial applications—ALD remains still under 5% of revenue.³²³ That baseline is essential context: while the technology functions and customers are buying it, converting initial industrial application into a business that moves the consolidated income statement has taken years and remains incomplete.

The competitive landscape in ALD is also far less forgiving than in PECVD. Market leader ASM International built its position over decades and holds the industry's deepest process library, alongside established offerings from Lam Research and Applied Materials. Because ALD films are deployed on mission-critical layers, customers are particularly cautious about swapping vendors. Winning a PECVD slot on a non-critical passivation layer is a fundamentally different engineering hurdle than winning an ALD slot on a DRAM capacitor dielectric. Consequently, Piotech's ¥301 million in 2025 ALD revenue must be evaluated with that distinction in mind—impressive as a growth rate, but early as a competitive position.³

Sub-segment two: hybrid bonding. Hybrid bonding represents Piotech's most ambitious strategic expansion. For fifty years, advancing chip performance meant making transistors smaller. As physical scaling grew prohibitively expensive, the semiconductor industry turned to 3D integration: placing two dies face to face and connecting them directly.

Traditional packaging joins stacked chips with micro-bumps—tiny solder balls that are relatively large, introduce electrical resistance, and limit total interconnect density. Hybrid bonding eliminates solder entirely. Two surfaces, each a mosaic of copper contact pads set within dielectric insulation, are polished to near-atomic flatness, brought into contact, and bonded at low temperatures: dielectric to dielectric first, followed by copper to copper as the metal expands under annealing. The result resembles welding two mirrors together so precisely that the seam disappears.

This yields interconnect densities orders of magnitude beyond what solder bumps allow, making hybrid bonding essential for high-bandwidth memory (HBM) stacks, stacked CMOS image sensors, and YMTC's Xtacking architecture, which fabricates the 3D NAND memory array and its peripheral logic on separate wafers before bonding them together.

The equipment market for hybrid bonding has been effectively monopolized by European suppliers: Austria's EV Group leads wafer-to-wafer bonding, the Netherlands' BE Semiconductor (Besi) leads die-to-wafer assembly, and SUSS MicroTec supplies complementary tooling, with core technologies protected by dense patent portfolios. Piotech Keytech entered this space by releasing a wafer-to-wafer hybrid bonding product line and a die-to-wafer surface preparation product, targeting a global market that Besi's neutral-case projections estimate at roughly 1,400 hybrid bonding systems by 2030.

Why would a state-backed fund pay 44 times book value for ¥1.15 million in half-year revenue? Because it was buying an option on securing domestic access to a tool category that gates advanced packaging and domestic HBM production. Evaluated as a strategic option, the valuation reflects policy priorities; evaluated as an equity stake in an operating business, it highlights the disconnect between current financial returns and pricing.

Two structural features of this market deserve emphasis. First, hybrid bonding is a packaging tool, meaning Keytech sells to a different customer set than its parent—advanced packaging houses and back-end memory operations—and competes against distinct incumbents. Second, the underlying economics remain unproven at Chinese production volumes. Bonding tools are high-precision alignment machines with sub-100-nanometer placement requirements, sold in small unit volumes at high unit prices, where low-volume manufacturing carries an unforgiving cost structure. A subsidiary generating single-digit millions of renminbi in half-year revenue against a ¥37 million half-year net loss is nowhere near the volume where any of that resolves.²¹

Customer concentration poses an additional hurdle. Wafer-to-wafer bonding is used by a small group of memory and image-sensor manufacturers, while die-to-wafer assembly serves advanced packaging facilities supporting AI accelerators. In China, both segments are dominated by a handful of firms, several of which face international trade restrictions. An equipment maker with an addressable customer list that can be counted on two hands inherits its parent's customer concentration risk in an acute form, without the parent's sixteen-year history of relationship-building.

Sub-segment three: the gap-fill specialties. HDPCVD and SACVD are the specialized tools that fill deep trenches. As chip structures become taller and narrower, filling them without creating internal voids becomes a major materials-science challenge. High-density plasma CVD simultaneously deposits and sputters material to keep trench openings from pinching shut, while sub-atmospheric CVD uses ozone chemistry to flow into deep features. Together with Flowable CVD, these specialty lines account for the remaining 14% of Piotech's revenue. Their primary significance lies less as an immediate growth narrative than as evidence that Piotech can adapt its core chamber platform across adjacent process types—a prerequisite for executing a broader platform strategy.

Testing the optionality claim. Bullish investment arguments frame ALD and hybrid bonding as near-term, high-margin growth drivers. However, Piotech's historical execution counsels patience. The company's demonstrated timeline from technical milestone to material revenue is extended: moving from the first industrial application of thermal ALD and HDPCVD to ¥301 million of ALD revenue took years, while hybrid bonding revenue of ¥136 million remains an entirely different order of magnitude from the ¥5.1 billion core.³²³

Crucially, demand for hybrid bonding tools is outside Piotech's direct control. It depends on domestic advanced-packaging capex, on whether Chinese HBM programs ramp on schedule, and on whether the industry's interconnect roadmap continues to favor hybrid bonding over competing approaches.

Management's disclosures reflect this cautious reality. At the June 2024 results briefing, Chairman Lv stated that the company would "continue to promote hybrid bonding equipment customer expansion, gradually deepen market understanding, and further explore applications" in memory, image sensors and advanced packaging.²³ Phrases such as "gradually deepen" and "explore" signal early-stage market development rather than imminent scale. Furthermore, when asked at the July 2026 briefing about the pending acquisition, overseas expansion, and new order volumes, management declined to elaborate beyond pointing to published announcements.¹⁹

The defensible conclusion is clear: Piotech's non-PECVD product lines represent genuine strategic options, intelligently funded by sharing capital intensity with state investment partners rather than absorbing it on the parent balance sheet. But those options remain unproven at commercial scale. Confirming their economic value will require not another technology announcement, but a financial year in which hybrid bonding revenue expands multi-fold from a base of hundreds of millions of renminbi, at transparent gross margins, with repeat orders from established customers.

How that funding was arranged, however, opens a separate and less comfortable line of inquiry into corporate governance and capital allocation.

↑ Back to Top

VII. Management, Governance & Capital Allocation

In September 2025, the same month Big Fund Phase III agreed to anchor Keytech, Piotech announced a combined financing package exceeding ¥5.6 billion across the parent company and its subsidiary—up to ¥4.6 billion through a private placement at the listed parent, and up to ¥1.04 billion at Keytech.²² The parent designated its proceeds for an industrialization base for high-end semiconductor equipment, a frontier technology R&D center, and additional working capital.²²

During the weeks surrounding that announcement, several key insiders reduced their holdings. Employee holding platforms sold nearly seven million shares (6.9932 million) in August 2025 at ¥144.92 per share, realizing approximately ¥1.013 billion.²² A group of senior executives—including the general manager, board secretary, four vice presidents, and the head of finance—subsequently disclosed plans to reduce equity incentive holdings between September 18 and December 17, 2025.²²

Such sales represent a standard regulatory mechanism for monetizing executive incentive compensation. Yet the sequence remains notable: management sought ¥4.6 billion in new growth capital from equity markets while key executives and employee holding platforms were selling into that same market.

The placement itself, and what it revealed about the market. The China Securities Regulatory Commission approved the registration on April 20, 2026, and the placement closed on June 22, 2026. Piotech issued 7.986 million shares at ¥576.01 per share, raising ¥4.6 billion gross and ¥4.558 billion net across twelve institutional investors—including E Fund Management, a Guotou group fund, and UBS AG—subject to a six-month lock-up.²⁴ Big Fund Phase III anchored the round with an ¥850 million investment.²⁵

That placement price of ¥576.01 per share stands in sharp contrast to the company's ¥71.88 IPO price in April 2022.¹³ Driven by a broader sector rally, Piotech's stock rose more than 150% in the first half of 2026 alone, pushing its market capitalization past ¥230 billion prior to a trading suspension tied to a major acquisition announcement.²⁵ Relative to full-year 2025 net profit of ¥927 million—a figure significantly boosted by non-operating gains—this valuation reflects aggressive market expectations. The public market has priced Piotech's emerging product pipeline under the assumption of rapid commercial conversion, even though historical evidence shows that tool adoption in semiconductor equipment unfolds over multi-year cycles.

The structure of Keytech's fundraising offers a clear rationale. By securing external capital at the subsidiary level, Piotech financed a capital-intensive R&D effort without diluting parent equity or fully absorbing early-stage operating losses. Retaining a 53.57% controlling stake allows the parent to consolidate Keytech's financials. If hybrid bonding achieves commercial scale, Piotech retains majority ownership of the upside; if development stalls, downside risk is shared with a state fund structured specifically for long-horizon industrial bets. As the firm's first major subsidiary-level financing, this arrangement establishes a precedent for risk-sharing, though its long-term effectiveness remains to be proven.

The ownership structure. Piotech has no single controlling shareholder. As of the third quarter of 2025, Big Fund Phase I was the largest shareholder with a 19.57% stake (55.0267 million shares), followed by a Guotou-affiliated Shanghai technology transfer fund at 13.54%.²⁶ The board of directors mirrors this institutional balance: alongside Chairman Lv and Director Jiang sit nominees from the Big Fund's management entity and the State Development & Investment Corporation, an independent director from the Chinese Academy of Engineering specializing in microelectronics, and Gerald Yin, chairman of AMEC.⁸ Executive management includes Liu Jing as legal representative and Yang Xiaoqiang as head of accounting.²⁷

Without a controlling shareholder, strategic governance relies on institutional consensus among stakeholders whose objectives—policy localization, capital returns, regional industrial growth, and supply-chain security—do not always align.

Furthermore, state capital positions are shifting. Between December 12, 2025, and March 11, 2026, Big Fund Phase I executed a plan to trim its stake by up to 8.435 million shares—representing 3% of Piotech's total equity—via block trades valued at approximately ¥2.597 billion, citing portfolio management requirements.²⁶ The fund announced a parallel stake reduction in Yandong Micro during the same period.²⁶ While Big Fund Phase III reinvested through the private placement and Keytech, Phase I's monetization demonstrates that state fund backing does not function as an unyielding price floor; early phase funds systematically recycle capital as portfolio assets mature.

Evaluating management guidance requires navigating constrained corporate disclosures. Piotech does not issue detailed, quantitative forward guidance typical of international peers, preferring annual briefing sessions over quarterly analyst Q&A calls. Key regulatory filings—such as responses to the Entity List designation and changes in core technical designations—have remained brief.

Where management assertions can be cross-checked against performance, results show alignment: management's explanation that 2025 margin pressure stemmed from new product verification costs was supported by a gross margin recovery to 41.69% in the first quarter of 2026.⁶¹⁹ However, on queries regarding overseas expansion, order backlog growth, or valuation terms for pending acquisitions, executives have routinely declined to provide detail beyond statutory announcements.¹⁹

Employee alignment relies heavily on equity incentive programs and holding platforms. In an organization without a controlling founder, retaining process engineers—who account for 726 of the company's 1,696 employees—represents a key operational priority, and equity serves as the primary instrument for it.³ These equity structures inherently generate periodic insider sales as vestings mature, creating selling pressure during periods of market strength.

Capital allocation. In 2025, R&D expenditure reached ¥859 million, representing 13.18% of annual revenue, with 2,140 cumulative patent applications and 707 granted patents, including 362 invention patents. R&D personnel comprised 42.81% of the total headcount, or 726 out of 1,696 employees.³ This staffing mix reflects an organization focused on expanding platform capabilities.

However, two aspects of its capital deployment warrant attention. First, spending intensity relative to revenue has moderated as scale has expanded: first-quarter 2026 R&D expense of ¥156.56 million represented 16.88% of revenue, down from 22.38% in the prior-year period.²⁸ While operational leverage naturally reduces R&D as a percentage of sales during revenue expansion, maintaining absolute R&D growth remains critical for closing technology gaps with global equipment leaders.

Second, dividend distributions remain modest. The payout ratio for 2025 stood at 10.02%. During the July 2026 results briefing, management responded to shareholder inquiries regarding cash returns by reiterating that capital retention is necessary to fund expansion projects and technical development.¹⁹

Balance sheet leverage has remained conservative. Total assets stood at ¥19.824 billion against shareholders' equity of ¥6.610 billion at the start of 2026, with customer prepayments making up the largest share of liabilities rather than interest-bearing borrowings.³ While prepayments represent delivery obligations rather than uncommitted cash, the absence of debt provides financial flexibility across industry cycles.

The Shangji acquisition. On June 26, 2026, Piotech announced plans to acquire a controlling interest in 无锡尚积半导体 Wuxi Shangji Semiconductor—a firm founded in 2021—via equity issuance and cash, with final valuation terms pending audit.²⁵ Shangji specializes in physical vapour deposition (PVD) sputtering systems and dry etch tools. Incorporating these lines alongside Piotech's CVD and ALD capabilities would create a combined deposition and etch equipment platform.²⁵ Shortly thereafter, in July 2026, Piotech committed ¥4.165 billion in capital injections across two wholly owned subsidiaries to accelerate manufacturing scale and technology development.

The strategic rationale aligns with industry trends toward multi-tool platforms, mirroring the expansion strategies of broader equipment vendors. However, this transaction represents Piotech's first major corporate acquisition, announced shortly after completing its ¥4.6 billion private placement. It targets an entity operating in PVD and etch—sectors where Piotech lacks an established operational history and where domestic peers Naura and AMEC hold mature positions. During the July 2026 briefing, Chairman Lv stated that the transaction remained in the planning and evaluation phase.¹⁹

Because Piotech has no prior M&A track record, its capacity to execute large acquisitions and integrate adjacent tool technologies remains untested. Key benchmarks for evaluating the acquisition will include the final transaction price, resulting goodwill, and whether Shangji's PVD and etch tools can secure independent qualifications on active fab lines.

↑ Back to Top

VIII. Strategic Analysis: 7 Powers, Porter's 5 Forces & Bull vs. Bear Stress Test

Strip away geopolitics and state support, and the central question for a long-term owner is straightforward: what structural barriers prevent Piotech from being competed down to commodity returns?

Hamilton Helmer's 7 Powers, applied honestly.

Switching costs — strong, but narrower than advertised. The mechanism is real: qualified recipes on active fab lines do not move. However, qualification is granted per tool and per process node. While it protects installed machines, it does not secure the next procurement tender. The margin contraction from 2023 to 2025 serves as an empirical test of whether these switching costs translate into pricing power; over that window, they did not.

Process power — strong, and the most defensible claim in the story. Multi-chamber vacuum architecture, plasma uniformity, thermal control across a 300mm wafer, and recipe libraries accumulated across 457 million wafer passes constitute proprietary know-how that cannot be easily bought or reverse-engineered.³ This is the competitive advantage most likely to endure. Its primary vulnerability is human: process power resides in engineers, and Piotech's founding cohort of US-trained process architects is in its sixties and seventies and has been progressively removed from formal core technical designations.¹⁸

Scale economies — moderate and improving. Piotech holds the largest domestic PECVD installed base to absorb R&D costs, and executive leadership attributed a portion of the first-half 2026 margin recovery to unit-cost reductions from production scale.⁴ Compared to Applied Materials' global scale, however, Piotech's absolute R&D budget remains small. Furthermore, scale in capital equipment relies heavily on service-network density—an area where Piotech benefits domestically but maintains virtually no international footprint.

Counter-positioning — moderate and time-limited. Piotech can dispatch a field engineer to a Wuhan fab on short notice. A US supplier operating under export restrictions may be legally prohibited from servicing that same machine. That asymmetry is a genuine operational advantage, but it remains entirely a product of trade policy. It provides no defense if regulatory conditions change, nor does it assist Piotech outside China.

Cornered resource, network effects, branding — weak to negligible. Fabs make purchasing decisions based on yield, cost of ownership, and delivery reliability. There is no brand premium in semiconductor capital equipment, nor are there network effects between customer lines.

Porter's Five Forces.

New entrants — low threat. A decade of operations and substantial cumulative losses before generating meaningful commercial revenue represent a formidable barrier to entry, as Piotech's own corporate history demonstrates.¹⁰¹¹

Buyer power — high, and the binding constraint. Procurement is concentrated among roughly four primary fabs, with the top five customers generating 79.15% of 2025 revenue.³ Coupled with an explicit national policy encouraging multi-sourcing and the active qualification of domestic alternatives, buyer power represents the primary constraint on Piotech's returns.

Supplier power — high to moderate, and partially self-remediated. Radio-frequency generators, mass flow controllers, vacuum pumps, and precision seals remain critical supply bottlenecks, exacerbated by Entity List restrictions on foreign sourcing.⁷ Piotech's strategy of taking equity stakes in key domestic suppliers, such as Hengyunchang, has reduced component dependency, though it entangles component costs with investment gains on its balance sheet.¹⁶¹⁷

Substitutes — low threat. Chips cannot be manufactured without thin-film deposition. The primary competitive risk is that an alternative deposition technique displaces a specific process layer, not that deposition as a discipline becomes obsolete.

Rivalry — intense and intensifying. Piotech faces incumbents Applied Materials and Lam Research for advanced-node process slots, alongside Naura's significantly larger multi-product equipment platform domestically.²⁰

The bull case, stated at its strongest. Domestic substitution in thin-film deposition remains in its early stages, with historical localization rates in single digits. Consequently, a company already generating ¥5.142 billion in PECVD revenue retains a substantial growth runway simply by displacing foreign tools in domestic fabs.

Forward order visibility remains exceptionally strong for an equipment vendor, supported by an ¥11 billion backlog and ¥4.852 billion in customer prepayments representing committed demand rather than speculative forecasts.³ The gross margin recovery during the first half of 2026 reinforces management's explanation that 2025 margin compression reflected product-transition costs rather than structural price erosion.⁶⁴

Additionally, secondary growth initiatives are being co-funded by external partners. ALD sales are expanding rapidly from a small base, hybrid bonding is supported by state equity at the subsidiary level, and the proposed Shangji acquisition—if completed on favorable terms—would transform a single-process specialist into a broader deposition-and-etch equipment platform. Having survived a loss-making decade to establish its initial product line, Piotech now possesses both the capital and customer access to scale adjacent categories.

The bear case, stated at its strongest. Gross margin remains well below its 2023 high of over 50%, and management has not guided to a full return to those levels. If gross margins stabilize in the low 40% range, the market is assigning a platform valuation multiple to a hardware vendor whose realized pricing remains subject to the bargaining power of four primary buyers.⁵⁶

Balance-sheet risks remain prominent. Total inventory of ¥7.826 billion represents more than half of current assets, with ¥4.362 billion sitting on customer sites awaiting formal acceptance—a category where filings explicitly cite impairment risk if sign-offs stall.¹² Operating cash flow, which was positive in 2025 due to customer prepayments, turned negative at ¥519.85 million in the first quarter of 2026, indicating that cash generation was driven by prepayment timing rather than a permanent structural shift.³²⁸

Reported net profit has also been elevated by non-operating items, with non-recurring gains accounting for over 82% of first-quarter 2026 net profit.²⁸ Heavy customer concentration leaves the company exposed to capex pauses at major fabs during memory downcycles or policy shifts. Furthermore, tighter component export controls could disrupt tool assembly before domestic supply chains achieve full operational reliability, while Naura's expansion into CVD from a larger platform base represents a direct domestic challenge. Meanwhile, anchor investor Big Fund Phase I has been reducing its equity position.²⁶

The risks that actually apply. Evaluating macro risks requires isolating mechanisms that directly impact Piotech's operating business.

Demand cyclicality represents the primary commercial vulnerability. Semiconductor equipment spending is derived from fab capital expenditure, which tracks chip pricing. Exposure to memory manufacturers means a pullback in domestic DRAM or 3D NAND expansion transmits directly into reduced tool orders, with customer concentration leaving limited revenue diversification to cushion the impact.³

Supply-chain and geopolitical risks operate counter to public narrative assumptions: tightening export controls on imported components represent a far more immediate threat to Piotech's production than restrictions on customer sales. Integration risk tied to the Shangji transaction poses a separate challenge, as Piotech has no prior track record of executing corporate acquisitions.

Refinancing risk remains low given recent equity raises and customer prepayments. However, a significant portion of apparent liquidity consists of customer prepayments tied to undelivered hardware. Finally, technology-disruption risk—while low across thin-film deposition generally—remains relevant for specific tool categories, where customer process redesigns can eliminate demand for individual tool lines.

The activist's questions, which an investor should want answered: What is the underlying gross margin by individual tool line, specifically for hybrid bonding? What portion of the ¥4.362 billion shipped-but-unaccepted inventory has been sitting at customer fabs for longer than twelve months? What final purchase price is being paid for Wuxi Shangji, and how much goodwill will be recorded? Given that Piotech is both an equity holder and the primary customer of listed supplier Hengyunchang, what are the exact pricing and payment terms governing those related-party transactions? And why, during a period in which the company raised ¥4.6 billion in new equity capital, were executive officers and employee platforms net sellers of shares?

None of these questions carry inherently damaging answers, but all remain unanswered in public disclosures.

↑ Back to Top

IX. Playbook, Key Investor KPIs & Epilogue

There is a photograph that does not exist but should: a Shenyang industrial park in the winter of 2012, a handful of engineers in an unheated building, a single PECVD chamber, and no customers. Set it beside a 2026 balance sheet carrying ¥7.8 billion of inventory and a market capitalisation in the hundreds of billions of renminbi, and the distance between those two frames tells the entire story. What is striking is that almost nothing across that span was driven by technological invention alone. It was accomplished through persistence, state patience, and a geopolitical shift that nobody in that building could have anticipated.

Four durable lessons.

Certification is not commercialisation. Piotech's first decade produced national R&D awards, patents, and operational 12-inch PECVD hardware — yet generated just ¥70.64 million in revenue in year eight alongside a net loss of ¥103 million.¹⁰ The moat in wafer fab equipment is not the tool itself; it is the qualified installed base, and bridging the gap between them requires years of validation and substantial capital. Investors should apply that same framework to every milestone announced for hybrid bonding, as well as every claim made regarding the Shangji etch and PVD lines.

The buyer's structure sets the ceiling on the seller's margin. Piotech's customers are few, large, strategically coordinated, and institutionally committed to maintaining at least two qualified suppliers per process step. In that market structure, technical excellence buys sales volume and operational survival; it does not confer pricing power. This represents the single most transferable lesson from the story, applying equally to any supplier operating in sectors like defence or infrastructure procurement where demand originates from concentrated buyers enforcing explicit second-sourcing policies.

Downstream tailwinds are bounded by upstream dependencies. US export controls simultaneously secured Piotech's customer demand and threatened its critical component supply. The company's strategic response — taking an equity stake in a domestic plasma RF power supply manufacturer that subsequently listed at a substantial valuation gain — represents its most agile operational move.¹⁶ Yet it also underscores an enduring reality: a supply chain built through vertical partnerships requires continuous capital commitment and creates long-term purchasing obligations.

A single-product equipment vendor must become a platform, and the transition is where value is destroyed. Naura's multi-product platform model represents the target Piotech is now pursuing through ALD development, hybrid bonding, and external acquisitions. Each move expands the company beyond the core PECVD capability that established its business, yet public market valuations currently price these expansion efforts as though successful commercial execution is assured.

An overarching lesson sits underneath these four: state capital functions not as a permanent subsidy, but as a rotating fund operating on a distinct return timeline. The state fund that anchored Piotech through its unprofitable decade has systematically reduced its position during market strength, while its successor fund deploys capital into next-generation assets. Investors who view state ownership as an unyielding price floor mistake a managed investment portfolio for an unconditional political guarantee.

Three KPIs that actually matter.

One: gross margin, quarter by quarter, against the 2023 benchmark. This single metric evaluates the central claim of the bull thesis — specifically, whether a qualified installed base produces pricing power or merely unit volume. Executive management stated that margins should stabilise following the first-quarter 2026 rebound.¹⁹ The key test is whether gross margins settle in the low 40% range or recover toward historical peaks near 50%.

Two: shipped-not-accepted goods (发出商品) relative to revenue, and any impairment charges against them. This line item serves as a combined indicator of cash-conversion efficiency and underlying demand quality. If shipped inventory grows more slowly than recognized revenue, the business model is maturing and cash generation is becoming structural. Conversely, if shipped goods continue outstripping revenue growth, volume is being financed via balance-sheet working capital, leaving the company exposed to direct inventory write-downs if customer acceptances stall.¹²

A note on what not to track. Cumulative chamber shipments and total production-line deployments are the metrics management highlights most prominently, providing clear evidence of installed-base reach. However, because these tallies are cumulative, they only move upward and reveal nothing about realized pricing, gross margins, or whether tools delivered in a given period were accepted and converted into cash. Cumulative milestones serve as effective public relations points, but offer limited utility as financial performance metrics.

Three: the mix of profit between operations and investment gains. Because non-recurring items generated the majority of reported net profit in the first half of 2026, net profit excluding non-recurring items — ¥367 million in the first half of 2026 and ¥723 million for full-year 2025 — represents the core metric evaluating the underlying equipment business.⁴³

These three key performance indicators are derived directly from published financial statements. Each carries an explicit test of the investment thesis: sustained gross margins below the low 40s would challenge the pricing-power thesis; a shipped-goods balance continuously expanding faster than revenue would undermine claims of capital efficiency; and a persistent gap between headline net income and operating profit would indicate that investment gains, rather than tool sales alone, are supporting current valuation levels.

Epilogue. Piotech occupies a unique position within China's semiconductor ecosystem: strategically essential to domestic supply security, yet operating as a publicly listed enterprise evaluated on financial returns. The company was established by US-trained engineers in a traditional industrial city, sustained through a loss-making decade by national R&D programs and state equity, provided a captive domestic market by US export controls, and subsequently designated on the Entity List by the same regulatory authorities whose restrictions accelerated its adoption. Its initial state anchor investor has been trimming its position into market strength, while a successor fund anchors its latest capital raises.

This dual identity is a structural feature of the asset. Piotech functions simultaneously as a commercial firm answerable to public shareholders and as an execution vehicle for national technology strategy. While those objectives frequently align during capacity expansion cycles, they can diverge. Decisions to maintain low-margin equipment lines for strategic process coverage, or to accept demanding commercial terms from key domestic chipmakers, may align with national policy objectives while diluting shareholder returns. Public disclosures provide limited visibility into how management navigates such competing priorities when trade-offs arise.

The company has successfully achieved the most demanding operational hurdle: manufacturing deposition tools capable of running continuous production volumes on advanced customer lines. Deploying approximately 3,800 process chambers across roughly 100 production lines demonstrates genuine industrial qualification rather than policy reliance.⁴ What remains unproven is whether this operational baseline can yield the financial profile of global equipment leaders — characterized by sustained 45%-plus gross margins, disciplined pricing, multi-product platform scale, and consistent cash conversion.

Evaluating Piotech requires avoiding premature conclusions in either direction. The optimistic and cautious cases reflect different analytical timeframes. Over the medium term, the domestic equipment substitution trajectory remains supported by ongoing fab construction commitments. Over a full industry cycle, however, the long-term economics of operating a deposition equipment maker in a market dominated by four primary buyers and multiple domestic vendors remain unproven, with the 2023-to-2025 margin trajectory offering the primary empirical record to date.

The coming two years will provide clarity. The gross margin trend will indicate whether recent margin pressure reflected a temporary product transition or a permanent structural repricing. The Shangji integration will establish the company's capital allocation track record in corporate M&A. Hybrid bonding adoption will determine whether secondary product lines achieve commercial scale or remain specialized R&D projects. And the balance of shipped, unaccepted inventory will either convert predictably into cash or continue to absorb working capital. Having spent its first decade proving it could build production-grade tools, Piotech's main test is whether those tools generate sustainable economic returns.

References

1. 拓荆科技：PECVD龙头看上去很美 — 第一财经 Yicai, 2022-05-20 ↩↩↩↩↩↩

2. 拓荆科技股份有限公司官方网站 — Piotech Inc. ↩↩↩

3. 拓荆科技2025年营收65.19亿元同比增58.87% 在手订单约110亿元支撑持续增长 — 新浪科技 Sina Tech, 2026-04-28 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

4. 拓荆科技炒股大赚归母净利暴增13倍 深耕高端产品进入100条半导体生产线 — 新浪财经 Sina Finance, 2026-08-24 ↩↩↩↩↩↩↩↩↩↩

5. 拓荆科技2025年报解读：扣非净利增103%现金流转正，毛利率下滑风险需警惕 — 搜狐财经 Sohu Finance, 2026-04-28 ↩↩↩

6. 拓荆科技上半年归母净利润同比增长1324.10% — 中国基金报 China Fund News, 2026-08-24 ↩↩↩↩↩↩↩

7. 实体清单新增136中国实体，半导体设备厂商密集受限 — 第一财经 Yicai, 2024-12-03 ↩↩↩↩

8. 拓荆科技 — 董事会及高级管理人员 Board of Directors and Senior Management ↩↩↩↩↩

9. 深耕半导体设备领域 拓荆科技跨越"0到N" — 中国证券报 China Securities Journal, 2024-12-06 ↩↩↩

10. 大基金为第一大股东 拓荆科技拟募10亿冲刺科创板 — 证券时报网 Securities Times, 2021-07-13 ↩↩↩↩

11. 拓荆科技股份有限公司首次公开发行股票并在科创板上市招股说明书 — 上海证券交易所 Shanghai Stock Exchange, 2022-04-06 ↩↩

12. 拓荆科技股份有限公司2025年度向特定对象发行A股股票募集说明书 — 上海证券交易所 Shanghai Stock Exchange, 2025-12-04 ↩↩↩↩↩↩↩

13. 历史高位股拓荆科技拟不超46亿定增 2022上市超募11亿 — 新浪财经 Sina Finance, 2025-09-15 ↩↩

14. Addition of Entities to the Entity List — Federal Register, 2024-12-04 ↩

15. 拓荆科技回应被列入"实体清单"：不对日常经营产生实质性影响 — 证券时报 Securities Times, 2024-12-03 ↩

16. 国产射频电源龙头恒运昌在科创板上市 — 中国证券报·中证网 China Securities Journal, 2026-01-28 ↩↩↩↩

17. 拓荆科技 恒运昌IPO的关键"双角" — 新浪财经 Sina Finance, 2025-11-12 ↩↩

18. 拓荆科技核心技术人员吕光泉离任 — 新浪财经 Sina Finance, 2024-12-14 ↩↩↩

19. 直击拓荆科技业绩会：先进产品已量产 — 21世纪经济报道 21st Century Business Herald, 2026-07-09 ↩↩↩↩↩↩↩↩

20. 北方华创科技集团股份有限公司2025年半年度报告 — 巨潮资讯网 CNINFO, 2025-08-29 ↩↩

21. 溢价超44倍！大基金三期为何青睐拓荆科技旗下子公司？ — 每日经济新闻 National Business Daily, 2025-09-13 ↩↩↩↩↩↩

22. 拓荆科技与子公司拟巨额融资超56亿元，高管集体减持同期实施 — 腾讯新闻 Tencent News, 2025-09-17 ↩↩↩↩↩

23. 拓荆科技：产品工艺开发及验证进展顺利 将继续推进混合键合设备客户拓展｜直击业绩会 — 财联社 Cailianshe, 2024-06-18 ↩↩↩

24. 拓荆科技：完成46亿元定增 发行价576.01元/股 — 财联社 Cailianshe, 2026-06-22 ↩

25. 拓荆科技：大基金三期加持，百亿订单下的并购棋局 — 新浪科技 Sina Tech, 2026-07-06 ↩↩↩↩

26. 大基金一期退出进行时：拟减持燕东微、拓荆科技，规模或超30亿 — 第一财经 Yicai, 2025-12-12 ↩↩↩↩

27. 拓荆科技股份有限公司2025年年度报告摘要 — 上海证券报 Shanghai Securities News, 2026-04-28 ↩

28. 拓荆科技2026年一季报解读：营收增56.97% 经营现金流净额大跌4858.29% — 新浪财经 Sina Finance, 2026-04-28 ↩↩↩