Shengyi Technology: The Copper-Clad Backbone of the AI & 5G Infrastructure Boom

I. Introduction & Episode Roadmap

In the summer of 2026, an unheralded industrial component supplier surged to the center of trading activity on the Shanghai Stock Exchange. 广东生益科技股份有限公司 Shengyi Technology Co., Ltd., which manufactures specialized copper-clad laminates, saw single-session turnover top RMB 10.5 billion in early August. After trading as low as RMB 40.40 over the preceding twelve months, the stock reached RMB 143.21 by mid-August 2026. The rally elevated Shengyi's market capitalization to roughly RMB 348 billion ($48 billion) after touching an intraday high of RMB 191.88.[^1]

While factory operations remained largely consistent, global demand shifted rapidly toward high-performance laminate materials essential for advanced computing hardware.

That core product is known as a 覆铜板 copper clad laminate (CCL), and Shengyi is among the few global suppliers capable of producing high-frequency, high-speed variants at scale. Listed on the 上海证券交易所 Shanghai Stock Exchange under ticker 600183, the company generated RMB 28.43 billion in 2025 revenue—a 39.45% increase—while net profit attributable to shareholders rose 91.75% to RMB 3.33 billion.¹² Its rigid CCL segment held approximately 13.7% of the global market in 2025, ranking second worldwide.³ Growth accelerated into 2026, with first-half net profit pre-announced at RMB 3.10 billion to RMB 3.30 billion—matching nearly its entire full-year 2025 profit in six months.⁴

The core hook. Founded in 1985 as a Sino-Hong Kong joint venture in Dongguan—when the surrounding Pearl River Delta was dominated by agriculture and light assembly—Shengyi has evolved into a critical element of global computing infrastructure. The company does not design semiconductors or operate silicon fabrication plants. Instead, its core capability lies in material formulation and process engineering: blending specialized resins, weaving fiberglass fabric, and bonding ultra-thin copper foil into laminates capable of carrying 224-gigabit-per-second signals across multi-layer circuit boards without severe signal degradation. In an AI server, laminate materials represent approximately one percent of total component costs yet can influence ten to twenty percent of usable system performance.⁵ This cost-to-performance asymmetry underpins both the investment thesis and the operational risks.

The strategic dilemma. Despite recent financial momentum, Shengyi faces significant competitive hurdles in high-value material slots. In 2025, Taiwan-based 台光电子 Elite Material Co. (EMC) surpassed both Shengyi and 建滔集团 Kingboard Holdings to become the leading global CCL supplier by revenue, capturing an estimated 16.2% of a market that expanded 24.4% to $18.68 billion.⁶ EMC achieved this position by securing primary material specifications in reference designs for North American AI accelerators and switches—the blueprints issued by chip designers and hyperscalers to circuit-board fabricators. Shengyi's strategic challenge is twofold: penetrate established design wins locked by competitors while defending gross margins against Kingboard, which maintains deeper backward integration into raw copper foil, resin, and glass fabric. Whether Shengyi can execute on both fronts amid volatile raw-material input costs remains the central business question.

Story roadmap. The analysis begins with the material physics and unit economics of copper-clad laminates, detailing why standard epoxy-glass sheets command RMB 12 while specialized high-speed laminates sell for over RMB 200. It traces Shengyi's founding in Dongguan and its transition away from commodity FR-4 laminates, followed by its expansion during the 5G deployment cycle and the 2021 listing of its printed-circuit-board subsidiary on the 科创板 STAR Market. Subsequent sections evaluate segment P&L drivers, the competitive dynamic in AI infrastructure materials from 2023 through 2026, structural positioning using competitive frameworks, management governance, and key risk metrics.

While recent financial results reflect a favorable market cycle, long-term performance remains tied to hardware demand cycles beyond the company's direct control, making rigorous operational scrutiny essential.

---

↑ Back to Top

II. The Physics & Economics of Copper Clad Laminates (CCL)

Picture a lamination press the size of a shipping container. Between its heated platens, workers stack alternating layers: sheets of woven glass fabric that have been soaked in liquid resin and partially cured to a tacky state, and on the outer faces, foils of electrodeposited copper thinner than a human hair. The press closes. Heat and pressure—several hundred degrees and tens of atmospheres—cross-link the resin into a rigid solid that permanently bonds the copper to the glass. Out comes a panel that looks like a copper-faced sheet of green fiberglass.

That panel is a copper-clad laminate, the foundation for every printed circuit board. Circuit board manufacturers etch away most of the copper layer to form signal traces, drill interconnect holes, and stack multiple laminates together using partially cured intermediate sheets—the 半固化片 prepreg—as the bonding agent. The laminate serves as the structural substrate, while the prepreg acts as the mortar. Shengyi manufactures both product lines and reports them within a single operating segment.

Why a boring material became a bottleneck

For most of the printed circuit board industry's history, a single material formulation dominated manufacturing: FR-4, a flame-retardant epoxy-glass composite. FR-4 is inexpensive, mechanically robust, and electrically sufficient for lower-frequency applications. However, at high signal frequencies, its physical limitations create significant performance bottlenecks.

Two primary electrical properties govern substrate performance. The first is the dielectric constant, Dk, which determines the velocity of signal propagation through the board—analogous to the optical density of a physical medium through which an electromagnetic wave travels. The second and more critical parameter is the dissipation factor, Df, which measures the proportion of signal energy absorbed by the material and converted into heat. In practical terms, Dk determines how much the material slows signal transmission, while Df measures signal attenuation along the trace.

At the switching speeds of 2005-era server platforms, FR-4's dissipation loss was negligible. At 112 gigabits per second—the PAM4 signaling rate standard in modern accelerator interconnects—and at 224Gbps, the generation entering system design, dissipation loss severely degrades signal integrity. Signals reaching the end of a 30-centimeter trace become attenuated and distorted, making it difficult for receiving chips to distinguish binary states. Every doubling of data transmission rates roughly doubles the attenuation penalty, compounding across every layer transition, microvia, and centimeter of trace length.

To address signal loss, the industry established a standardized hierarchy of material grades: standard loss, mid loss, low loss, very low loss, and ultra-low loss. In supply chain terminology, these correspond to grades M4, M6, M7, M8, and M9, each defining progressively lower dissipation targets. M6 and M7 materials operate in the 0.002 to 0.003 Df range; M8 targets 0.0012 to 0.002 and serves as the primary substrate for current-generation AI accelerator hardware; M9 achieves 0.0007 to 0.0009—representing a loss reduction of roughly 40% compared to M8—and typically requires replacing standard E-glass fabric with quartz or specialized low-Dk weaves.⁵ While M8 materials comfortably support 112Gbps architectures, M9 is required to make 224Gbps system designs commercially viable.⁵

Advancing up this material hierarchy requires fundamental formulation changes rather than minor process tweaks. Each grade demands distinct resin chemistries—such as polyphenylene ether systems, hydrocarbon blends, or PTFE—alongside specialized glass weaves, low-profile copper foil surface treatments, and tighter press parameters. Higher-performing grades also introduce manufacturing trade-offs: low-loss resin formulations frequently exhibit weaker mechanical adhesion to copper, reduced thermal cycling tolerance, and increased tool wear during mechanical drilling. Consequently, a laminate supplier's competitive differentiation rests not only on formulating low-Df compounds, but on producing them with consistent manufacturing yields at scale.

The cost anatomy, and why margins swing

A standard copper-clad laminate's bill of materials is approximately 40% to 45% copper foil, 25% to 30% resin system, and 15% to 20% electronic glass cloth, with energy, direct labor, and manufacturing overhead accounting for the remainder. This cost structure drives significant earnings volatility across industry cycles. Copper is an exchange-traded commodity subject to immediate price fluctuations; epoxy and specialty resins track petroleum derivatives and precursor chemical capacity; and electronic glass cloth depends on capital-intensive furnaces that require long lead times to construct.

Conversely, laminate selling prices reset on fixed contract cycles with circuit board fabricators, typically on a monthly or quarterly basis. For commodity laminate producers, gross margin performance operates largely as a spread trade defined by the lag between input cost changes and finished product price adjustments. When raw copper prices fall, gross margins expand temporarily until competitive pricing realigns margins. When input costs rise, margins compress until updated price schedules take effect across customer accounts.

The raw material environment in 2026 highlighted these margin dynamics. Electronic glass cloth prices roughly doubled from their third-quarter 2025 trough; ultra-fine electronic yarn increased from approximately RMB 25 per kilogram to double that level; lead times for high-end HVLP copper foil—required for low-profile, low-loss applications—extended into 2027; and select PPE resin formulations experienced threefold price increases following Middle East supply disruptions.⁷ Volume manufacturer Kingboard issued seven price increase notices in the first half of 2026 alone, including a 15% increase on June 16 coming just three weeks after its previous adjustment, resulting in a cumulative price increase exceeding 50%.⁷

Under typical market conditions, a rapid rise in input costs would compress laminate manufacturing margins. Instead, Shengyi's gross margin expanded to 25.10% for full-year 2025—up 4.52 percentage points year-over-year—and reached 28.1% in the first quarter of 2026, representing an increase of more than 14 percentage points from the prior-year period.¹⁸ This margin expansion occurred because high-spec laminates do not follow commodity pricing dynamics. When a specific material is designated in an accelerator reference design with only two or three qualified global suppliers, cost increases are passed through rapidly, protecting supplier spreads through capacity constraints rather than contract lag. Shengyi's margin performance during a major raw material cost increase demonstrates a structural shift in product mix toward higher-value specifications.

The analytical caveat is that pricing power derived from temporary supply tightness differs from pricing power sustained by structural competitive moats; supply-driven pricing power contracts as industry manufacturing capacity expands.

Understanding how a Dongguan-based joint venture established a position in high-specification laminates requires examining four decades of operational and industrial development.

↑ Back to Top

III. Origins & Corporate Evolution: From Dongguan Joint Venture to Global #2 (1985–2017)

In 1985, Dongguan was still a county of agricultural land and processing workshops, forty minutes by road from the Hong Kong border and heavily dependent on cross-border capital. Shanghai-born textile industrialist 唐翔千 Tang Hsiang-Chien, who had rebuilt his family's business empire in Hong Kong, emerged as one of the earliest cross-border investors in mainland Chinese manufacturing.

Tang anticipated that China's emerging electronics assembly hub would require vast volumes of printed circuit boards, which relied entirely on imported laminate sheets. To capture this demand locally, he established a Sino-foreign joint venture combining Hong Kong Weihua Electronics with Dongguan's municipal electronics corporation and a provincial foreign-trade entity to build a laminate facility.⁹ Annual production initially reached several hundred thousand square meters.¹⁰ By comparison, Shengyi sold 160.6 million square meters of copper-clad laminates in 2025.³

This joint venture ownership model provided strategic advantages beyond manufacturing technology. In 1985, Sino-foreign joint ventures gained access to imported capital equipment, international resin formulation licenses, and commercial governance standards unavailable to purely domestic state enterprises. Hong Kong Weihua Electronics remains a core investor, retaining a 12.44% equity stake—representing a four-decade institutional continuity rare among mainland industrial firms.¹¹

Listing, renaming, and the long apprenticeship

In 1998, Shengyi became the first Chinese laminate manufacturer to list on the Shanghai Stock Exchange, before adopting its current corporate name in 2000.⁹ While public capital funded automated production lines, the listing's broader impact was organizational: it subjected a state-linked joint venture to public disclosure standards and financial discipline well ahead of its domestic peers.

During the 2000s, Shengyi operated primarily in standard FR-4 laminates alongside numerous domestic competitors. The market was increasingly dominated by Kingboard, which leveraged extensive backward integration into raw copper foil, electronic glass fabric, and resin to operate at lower unit production costs.

For non-integrated producers, competing purely on price against a vertically integrated market leader presented severe margin risks. Without the capital or chemical expertise required to acquire raw-material supply chains upstream, Shengyi's executive team—led from 1990 by general manager 刘述峰 Liu Shufeng—focused on repositioning the business toward higher-specification segments where customer decisions depended on technical qualification rather than base material pricing.

This strategy required climbing the product ladder through incremental, defensible steps. Each transition obliged customers to complete rigorous material re-qualification, creating switching friction once specifications were locked.

The initial phase centered on environmental regulatory shifts. When European Union directives restricted hazardous substances in the mid-2000s, printed circuit board assembly transitioned to lead-free soldering processes operating at higher thermal peaks. To prevent board delamination and thermal cracking, fabricators adopted high-Tg laminates with elevated glass transition temperatures. Concurrently, European and Japanese original equipment manufacturers began requiring halogen-free flame-retardant formulations. Executing these compliance transitions favored manufacturers with internal resin chemistry capabilities over cost-focused assemblers.

The second phase targeted high-reliability industrial end markets. Automotive electronics, industrial automation, and telecommunications infrastructure required stringent validation cycles, evaluating thermal cycling tolerance, humidity resistance, and extended field durability. Although securing these approvals demanded multi-year testing programs with minimal immediate sales, successful qualifications locked in long-term customer relationships, as substituting substrates in automotive or industrial control units required high-level engineering approval.

A third critical factor was scaling process consistency across geographic regions. Transitioning low-loss formulations from lab prototypes to high-volume commercial production required precise operational controls. As Shengyi established manufacturing hubs across Dongguan, Shaanxi, Jiujiang, and Changzhou, management standardized resin recipes and pressing parameters into strict process windows, maintaining product yield consistency across multi-site production facilities.

By the mid-2010s, this multi-decade specialization had transformed Shengyi from a localized laminate plant into an R&D-intensive materials manufacturer. Over Liu Shufeng's 34-year tenure, the business expanded from an initial annual output of 600,000 square meters into an industrial group employing over 10,000 workers with net assets approaching RMB 14 billion.¹⁰

Throughout the 2010s, these strategic investments yielded steady rather than exceptional return on equity, as financial gains remained constrained by ongoing commodity market exposure. The commercial payoff for Shengyi's high-spec resin library depended on a broader telecommunications infrastructure upgrade—a catalyst that arrived with the global rollout of 5G networks in 2018.

↑ Back to Top

IV. The 5G Telecom Boom & The STAR Market Spin-Off (2018–2022)

Commercial momentum for industrial component suppliers often hinges on structural shifts in customer demand. For Shengyi, China's massive fifth-generation telecommunications deployment transformed years of material research into rapid top-line expansion.

Inflection point one: the base station surge

The technical requirements of 5G infrastructure altered substrate economics. Sub-6GHz active antenna units integrate dozens of transmit-receive channels into a compact enclosure, requiring low-loss signal routing at operating frequencies where standard FR-4 laminates fail. Antenna assemblies demanded specialized high-frequency materials, while baseband and backhaul systems required high-speed laminates. Equipment vendors such as 华为 Huawei and 中兴 ZTE drove base station installations across China at an unprecedented scale.

Concurrently, tightening U.S. export controls from 2019 onward prompted Chinese telecom original equipment manufacturers to minimize foreign supply dependencies. Equipment builders that previously relied on Japanese or Taiwanese laminate suppliers faced internal mandates to qualify domestic alternatives. Shengyi's multi-year investments in resin chemistry positioned the firm to fulfill these domestic sourcing requirements.

This dynamic altered the company's product mix. Revenue expanded alongside gross margins, securing a market position in high-specification categories where material qualification took precedence over unit price. Return on equity reached 21.6% in 2021—the highest level of the preceding decade.[^1]

A sharp cyclical contraction followed the initial build-out. Base station construction represents upfront capital expenditure rather than recurring software revenue; once core networks were established, procurement slowed. Combined with a post-pandemic inventory correction across consumer electronics, Shengyi experienced two consecutive years of declining financial performance. Revenue fell from RMB 18.01 billion in 2022 to RMB 16.59 billion in 2023, while net profit dropped from RMB 1.53 billion to RMB 1.16 billion.¹⁷ Return on equity compressed to 8.32%, approximately one-third of its 2021 peak.[^1] Industry participants, as noted in company disclosures, engaged in aggressive price discounting to maintain factory utilization.¹⁷

This downcycle highlights the operational leverage inherent in laminate manufacturing when end-market demand cools. Although Shengyi maintained overall profitability, sustained dividend payments, and uninterrupted R&D funding, net profit contracted 24% in a single year.

The period also underscored the nature of Shengyi's initial 5G market gains. Market share growth was assisted by geopolitical supply-chain policies alongside product performance, establishing a conditional advantage as the industry transitioned toward global AI infrastructure requirements.

Inflection point two: carving out the board business

Shengyi also addressed a structural tension in its corporate architecture: the ownership of an internal printed circuit board manufacturer.

Its subsidiary, 广东生益电子股份有限公司 Shengyi Electronics Co., Ltd., manufactured high-layer-count circuit boards for servers, telecom switches, and industrial equipment. While operationally sound, the subsidiary created commercial friction with independent board fabricators such as 深南电路 Shennan Circuits and 沪电股份 Wus Printed Circuit. Because these fabricators represented Shengyi's primary laminate customers while competing directly with Shengyi Electronics, purchasing raw materials from Shengyi indirectly funded a competitor's downstream capabilities.

Furthermore, high-layer circuit board production requires capital-intensive drilling and plating equipment, yielding structurally lower returns than laminate manufacturing. Funding these capital requirements from the parent company's balance sheet absorbed cash flow generated by core materials operations.

To resolve these capital and commercial pressures, Shengyi Electronics executed an initial public offering on the 科创板 STAR Market under ticker 688183 on February 25, 2021. The subsidiary issued 166 million shares at RMB 12.42 per share, raising RMB 2.066 billion in gross proceeds and RMB 1.975 billion net.¹² The transaction represented the first corporate spin-off of an A-share subsidiary onto the STAR Market under updated domestic listing regulations. Net proceeds funded high-speed board capacity at the Dongcheng facility, a second construction phase at Ji'an, and an R&D center.¹²

The transaction achieved three distinct objectives. First, downstream capital expansion was financed via external equity rather than parent cash flow. Second, the subsidiary secured an independent public equity vehicle, which enabled a subsequent RMB 2.6 billion proposed private placement in November 2025.[^14] Third, the structural separation improved commercial alignment with independent circuit board manufacturers.

However, because the parent company retained a 62.93% controlling stake in Shengyi Electronics, customer channel conflict was mitigated rather than fully eliminated.⁸ Independent fabricators remained aware that laminate purchases continued to support a majority-owned competitor. While consolidating a listed subsidiary preserved downstream manufacturing exposure during demand expansions, it also retained structural integration complexities across broader industry cycles.

↑ Back to Top

V. Core Business Breakdown & Segment Economics

Underneath its consolidated results, Shengyi operates two distinct primary businesses alongside a third emerging unit that functions primarily as long-term optionality.

The laminate core

In 2025, rigid copper-clad laminates and prepreg generated RMB 17.77 billion in revenue—a 20.17% increase year-over-year—with gross margin expanding 2.39 percentage points to 23.91%.¹ Shipment volumes offer critical insight into this trajectory: production rose 10.03% to 158.13 million square meters, while sales volume grew 11.95% to 160.63 million square meters.³ Because revenue expanded roughly eight percentage points faster than physical volume, average realization per square meter improved. For laminate manufacturers, revenue growth outpacing unit volume indicates a favorable shift toward higher-priced, advanced material specifications rather than pure volume expansion.

Manufacturing is distributed across production hubs in Dongguan, Shaanxi, Jiujiang, and Changzhou, providing regional proximity to major electronics fabricators alongside operational redundancy. Segment profitability varies sharply across product tiers: standard FR-4 laminates typically generate gross margins in the low-to-mid teens, high-speed and high-frequency grades yield substantially higher returns, and tight supply for ultra-low-loss materials elevates margins further. Consequently, the blended 23.91% segment margin reflects product mix composition, with financial performance determined primarily by the proportion of advanced substrate grades sold.

The board business

The printed-circuit-board subsidiary, Shengyi Electronics, generated RMB 9.14 billion in 2025 revenue—more than doubling with a 103.93% increase—while gross margin expanded 9.18 percentage points to 28.61%.¹ Momentum accelerated into the first half of 2026, as the subsidiary reported RMB 5.78 billion in revenue (up 53.46%) and RMB 1.11 billion in net profit (up 109.36%), supported by a 24.73% year-over-year increase in manufacturing floor area.¹³

This performance reverses the historical dynamic within the group. The downstream circuit board division—traditionally a capital-intensive, lower-margin segment—outperformed the laminate core on gross margin in 2025 while expanding revenue at five times the rate of the core business. This shift reflects market tightness in high-layer-count and high-density-interconnect (HDI) boards driven by AI compute infrastructure demand. Shengyi Electronics disclosed technical progress in five-to-six-stage high-order HDI and multi-layer boards exceeding thirty layers, confirming that these products received certification from major global computing platforms and transitioned to volume shipment.¹³ To support this output, the subsidiary is executing a multi-site expansion: scaling production at Dongcheng, completing initial operations at Ji'an, installing equipment for trial runs in Thailand, and commencing construction on a dedicated AI facility in Dongguan for advanced HDI production.¹³

The Thailand expansion reflects a broader strategy among domestic component suppliers to mitigate geopolitical trade friction and rules-of-origin requirements for Western supply chains, requiring substantial capital investment to insulate international revenue against shifting trade policies.

The optionality tier

挠性覆铜板 Flexible CCL (FCCL) and specialty substrate materials account for a minor share of total revenue. Flexible laminates utilize thin polyimide or polyester film substrates bonded to copper foil, providing mechanical flexibility for foldable smartphones, camera modules, and compact consumer electronics. The global FCCL market remains largely controlled by established Japanese and Korean suppliers, where Shengyi maintains a limited commercial presence.

Concurrently, Shengyi reported R&D progress in flip-chip ball grid array (FC-BGA) packaging substrate materials, low-dielectric HDI laminates, and low-carbon CCL variants.³ High-performance 封装基板 IC substrate materials command the highest gross margins in the laminate industry, but the sector remains highly concentrated: Resonac, Mitsubishi Gas Chemical, and Doosan collectively controlled over 80% of the market in 2024, according to Prismark data.¹⁴ Securing design wins in IC substrates would expand Shengyi's addressable market, but penetrating this oligopoly requires navigating multi-year qualification cycles against incumbents backed by extensive operational reliability records.

What the consolidated numbers say

At the consolidated level, gross margin reached 25.10% in 2025—an increase of 4.52 percentage points—while earnings per share rose 87.84% to RMB 1.39.¹ Operating cash flow expanded 262.17% to RMB 5.27 billion, outpacing net income growth and strengthening balance sheet liquidity.¹ Consolidated return on equity recovered to 19.94%, up from 11.67% in 2024 and an 8.32% trough in 2023.[^1]

Research and development expenditures rose 25.37% to RMB 1.45 billion, representing 5.10% of total revenue, with 2,031 R&D personnel comprising roughly 14% of the company's workforce.¹ Annual R&D spending has remained consistently within 4.8% to 5.7% of revenue since 2020, persisting across cyclical contractions.[^1] Given Shengyi's reliance on material formulation chemistry, sustaining R&D investment during the 2023 industry downturn—when net profit contracted 24%—helped preserve its technical development timeline for next-generation material specifications.

Working capital metrics present an area of ongoing operational exposure. Days of sales outstanding (DSO) stood at 141.9 days in 2025, improving from 161.9 days in 2024 but indicating average collection periods of roughly four and a half months.[^1] The company's cash conversion cycle measured approximately 163 days.[^1] Furthermore, first-quarter 2026 financial analysis highlighted accounts receivable reaching 294.56% of net profit, underscoring significant working capital tied up in trade credit.⁸ While expanding receivables are common during industry volume expansions, high receivables balances increase credit risk if downstream customer demand contracts.

Because in 2026, downstream hardware demand is dominated by a single primary driver.

↑ Back to Top

VI. The AI Server Shift & The Battle for High-Speed Materials (2023–2026)

Consider what an AI training rack demands of a circuit board. Dozens of accelerators communicate continuously and simultaneously at bandwidths that were considered backbone-network-scale a decade ago. The interconnect fabric linking these chips is not composed of external cables; for most of its length, signal routing depends on copper traces etched into printed circuit boards, midplanes, and switch cards. Those traces operate at 112Gbps in current deployments and 224Gbps in next-generation designs undergoing tape-out.

The financial and operational implications for materials suppliers are non-linear. An AI server consumes an estimated five to seven times the high-end CCL content of a conventional server.⁷ Meanwhile, the PCB content value of a single high-end AI system rose from roughly $35,100 to $116,700—a 233% increase.⁷ Multiplied across global data center deployments, this expansion created a severe demand shock, transforming a mature materials industry into a primary supply bottleneck.

Industry tracking from Prismark illustrates the magnitude of this expansion. High-speed CCL sales reached $4.18 billion in 2024—up 50.4% from $2.78 billion in 2023—within a rigid CCL market of $15.0 billion that grew 17.9%.¹⁴ By 2025, the overall market expanded 24.4% to $18.68 billion.⁶ The AI-specific segment of the CCL market measured roughly $2.2 billion in 2025, doubling year-over-year—a figure Shengyi cited to justify its largest-ever capital investment in capacity.²

The Taiwanese hegemony problem

At this stage, the competitive landscape presents significant challenges for mainland manufacturers.

Substrate material selection for North American AI accelerators and high-end switches is not determined by board fabricators. Instead, it is locked upstream in reference designs—the precise engineering specification packages issued by chip designers and hyperscalers to their contract manufacturing partners. These blueprints define the qualified laminate stack-up by supplier name and part number. Securing inclusion in a reference design requires twelve to eighteen months of rigorous signal integrity testing, thermal evaluation, and manufacturing trials. Once specified, incumbents are difficult to displace, as substitution forces customers to re-validate hardware without gaining performance advantages.

Taiwan-based Elite Material Co. (EMC) recognized this dynamic early and executed aggressively. By 2026, market analysis indicated that EMC secured an estimated 80% share on AWS Trainium, near-total allocation on Meta's MTIA platform, approximately half of Google's TPU requirements, and 50% to 60% of 800G and 1.6T switch platforms.¹⁵ These allocations represented locked specification wins rather than simple volume market share. 台燿科技 Taiwan Union Technology Corp (TUC) established a complementary position by qualifying its M9-grade substrate for next-generation midplanes alongside EMC and Panasonic on the Trainium roadmap.¹⁵

The global ranking shifted accordingly. In 2025, EMC captured the top spot in global CCL revenue from traditional leaders strictly through its high-margin AI product mix.⁶ A domestic industry leader holding a 13.7% global market share was overtaken by a specialized Taiwanese competitor with a shorter operating history in the category, because the specialist secured primary design slots.³⁶

Shengyi's counter-attack, and what is actually verified

Shengyi responded by targeting the specification layer directly, and by mid-2026, market evidence pointed to initial progress, though the details warrant careful scrutiny.

Industry reports in mid-2026 named Shengyi as the sole mainland Chinese manufacturer to complete Nvidia's M9-grade qualification, with validation reported in June 2026, while also describing its M8-grade material as securing primary share in a major North American customer's switch hardware.⁵⁷ If fully realized, this represents a critical strategic milestone: M9 substrates support 224Gbps architectures, and securing qualified supplier status ahead of a major platform ramp carries far higher economic value than expanding volume in commodity material grades.

However, company disclosures remain cautious. When asked directly on an investor interaction platform on July 6, 2026, whether it had submitted M10-grade samples to Nvidia and what stage certification had reached, Shengyi declined to provide specific details, stating only that it was "conducting a series of project collaborations on AI-related applications with major domestic and international end customers" and referring investors to official periodic filings.¹⁶ Addressing 2026 capacity allocation across M7, M8, and M9 grades, management noted that plant equipment possessed high operational flexibility and that final allocation would depend on prevailing market demand and pricing schedules.¹⁶

This reserved posture reflects both commercial prudence and standard non-disclosure obligations. For investors, however, it highlights a distinction between market speculation and regulatory disclosures. The reported M9 qualification remains an unconfirmed industry claim rather than a verified fact in official securities filings. Capital markets pricing Shengyi's equity at a RMB 348 billion valuation must recognize that a central element of the growth narrative relies on unverified external reports that await confirmation in upcoming periodic disclosures.

Nevertheless, financial metrics align with expanding international traction. Export revenue surged 102.82% in 2025, significantly outpacing domestic revenue growth of 26.65%.¹ Because global AI accelerator supply chains operate primarily in overseas markets, a doubling of export revenue strongly suggests broader participation in international hardware builds.

Myth versus reality: three things the market believes

Separating prevailing market consensus from verified evidence clarifies the company's fundamental position in mid-2026.

Myth: Shengyi is an AI company. Reality: Shengyi is a diversified materials manufacturer with an expanding AI segment whose precise revenue contribution remains undisclosed. Corporate filings group rigid laminates and prepreg into a single reporting line and printed circuit boards into another, omitting the specific revenue share derived from ultra-low-loss material grades. While the spread between total revenue growth and physical volume growth serves as a useful proxy for high-spec product mix expansion, it remains an indirect indicator.¹³

Myth: The M9 qualification confirms Shengyi as the dominant domestic AI materials supplier. Reality: While the reported M9 qualification is strategically significant, the company has neither confirmed nor denied the certification in official regulatory filings.⁵¹⁶ Furthermore, obtaining qualification represents an entry credential rather than guaranteed commercial volume. Final allocation on commercial platforms depends on unit pricing, production capacity, customer dual-sourcing strategies, and risk management policies—factors outside the supplier's sole control.

Myth: The transition to AI materials eliminates operational cyclicality. Reality: AI infrastructure investment is driven by capital expenditure from a concentrated group of hyperscale technology firms—a customer structure historically subject to sharp cyclical swings. While shifting from consumer electronics and telecommunications to AI infrastructure improves average margin quality, it exchanges one demand cycle for another rather than insulating the business from broader capital spending fluctuations.

High-speed networking represents a fourth key application that shares similar material requirements with AI servers. Transitioning network architectures from 400G to 800G and 1.6T switching—alongside associated optical transceivers—demands strict signal integrity, making switch hardware a major driver of high-end CCL demand.⁶ Shengyi's reported M8 design wins in a major North American customer's switch platform fall into this category.⁵ For investors, networking refresh cycles provide meaningful diversification against accelerator deployment schedules.

The domestic substitution vector

Domestic supply chain localization (国产替代 domestic substitution) represents the second core demand driver, offering greater certainty but lower gross margins than direct export sales.

Mainland technology enterprises—including 华为 Huawei, 百度 Baidu, 腾讯 Tencent, and 阿里巴巴 Alibaba—are expanding AI infrastructure under stringent U.S. export controls, compelling them to eliminate foreign component dependencies across their hardware stacks. This policy imperative creates captive demand for domestic materials suppliers capable of producing qualified high-speed substrates, positioning Shengyi as a primary domestic beneficiary.

However, domestic demand remains constrained by the total volume of Chinese AI silicon production. If domestic accelerator manufacturing faces bottlenecks—whether from semiconductor lithography limits, high-bandwidth memory (HBM) supply constraints, or fabrication yields—materials procurement slows accordingly. In its 2025 annual report, Shengyi explicitly cited geopolitical supply chain decoupling and memory component shortages among its key operational risk factors.¹ Consequently, while the domestic substitution trend provides a stable market floor, its growth trajectory remains tied to upstream semiconductor supply constraints beyond the company's operational control.

By mid-2026, Shengyi operates along two distinct strategic vectors: competing to break Taiwanese incumbents' reference design locks in international markets while supplying a protected, capacity-constrained domestic ecosystem. Evaluating whether that constitutes a winning long-term position requires examining the structural dynamics of the broader industry.

---

↑ Back to Top

VII. Competitive Landscape: Porter's 5 Forces & Hamilton Helmer's 7 Powers

                       GLOBAL RIGID CCL MARKET LANDSCAPE (2025)
+-----------------------------------------------------------------------------------+
|  1. Elite Material Co. (2383.TW)    ~16.2% Share  [AI Reference Design Lock]      |
|  2. Shengyi Technology (600183.SS)  ~13.7% Share  [Formulation Depth / Scale]     |
|  3. Kingboard Laminates (1888.HK)   Top-4         [Vertical Integration / Cost]   |
|  4. Nan Ya Plastics (1303.TW)       Top-4         [Integrated Chemical Scale]     |
|  5. Taiwan Union Tech (6274.TW)     Specialist    [High-Speed Server Focus]       |
+-----------------------------------------------------------------------------------+
   Top four suppliers ≈ 49.7% of the global market. Asia ex-Japan ≈ 96% of area
   consumption; China alone ≈ 74.8% of global value.

Market accounting from Prismark highlights an industry that is simultaneously consolidated and fragmented: four dominant suppliers control nearly half the global market, while hundreds of regional fabricators compete for the remainder.¹⁴⁶³ Yet this top-level framing masks a critical structural reality: the copper-clad laminate sector effectively functions as two distinct markets with fundamentally different economic characteristics.

Porter, applied honestly

Threat of new entrants. In commodity FR-4 laminates, entry barriers are minimal: production equipment is widely available, resin chemistry is public, and regional manufacturers add capacity continuously. In ultra-low-loss AI materials, however, entry barriers are formidable. Developing a high-specification laminate requires an estimated three to five years of research and development, followed by equipment manufacturer certification cycles that add another year or more.⁷ A new entrant attempting to commercialize M9-grade substrates today would arrive after the current AI architecture generation has been designed, deployed, and depreciated. This structural bifurcation represents the central economic reality of the industry.

Supplier power. Supplier leverage remains high and serves as the primary operational constraint across the supply chain. High-end HVLP copper foil is supplied by roughly three manufacturers controlling an estimated 80% to 90% of global capacity.⁷ Similarly, low-dielectric and quartz glass fabrics are produced by a concentrated group of Japanese and Chinese suppliers, while specialty polyphenylene ether (PPE) and hydrocarbon resins rely on a limited number of qualified chemical vendors. Because Shengyi depends on oligopolistic raw-material suppliers, its output growth is fundamentally constrained by upstream capital expenditure decisions. This supplier concentration underscores the strategic value of competitor Kingboard's deep backward integration.

Buyer power. Buyer power is structurally moderate but currently weak. Direct customers are printed circuit board fabricators, including Shennan Circuits, Wus Printed Circuit, and 胜宏科技 Victory Giant Technology. However, in high-specification material tiers, fabricators do not select laminate substrates independently; instead, they build to strict stack-up specifications dictated upstream by chip designers and hyperscalers, neutralizing fabricator bargaining power on qualified materials. In 2026, with customer lead times extending to two to three months and allocation protocols in force, buyer power collapsed almost entirely.⁷ However, this leverage will normalize as supply constraints ease.

Threat of substitutes. Alternative substrate technologies represent a meaningful long-term risk. In its 2025 annual report, Shengyi explicitly listed emerging ceramic and glass-based substrate technologies among its identified risks.¹ Glass-core substrates have drawn substantial capital investment from major semiconductor packaging firms seeking higher interconnect density. While commercial adoption faces long engineering timelines and manufacturing hurdles, official corporate risk disclosures signal that alternative substrate architectures could challenge traditional organic laminates over time.

Rivalry. Industry rivalry is intense and asymmetric. In commodity laminates, competition is price-driven and margin-dilutive. In high-speed AI grades, competition operates as a winner-take-most qualification race: securing a reference design slot locks in revenue for an entire hardware platform generation, while unselected vendors must wait eighteen months for the next architectural window.

The 7 Powers reading

Process Power represents Shengyi's primary competitive moat, developed across four decades of manufacturing operations. Achieving consistent, high-yield production of low-loss laminates at scale depends on proprietary resin blending protocols, thermal press calibration, glass cloth handling, and institutional process expertise rather than patented chemical formulas alone. Shengyi's ability to expand gross margins during severe raw-material cost inflation demonstrates this operational execution. At the end of 2025, the company held 435 valid patents, with 38 granted during the year, reflecting an extensive underlying foundation of trade secrets and process engineering know-how.¹

Switching Costs provide strong protection once material specifications are locked. After a printed circuit board design is finalized for a specific laminate stack-up, substituting material suppliers requires re-running extensive signal integrity and thermal validation tests. This creates substantial switching friction that protects incumbent suppliers. This dynamic reinforces EMC's dominant position in North American AI platforms, while simultaneously insulating Shengyi's established positions in domestic telecommunications and industrial hardware. Consequently, switching costs defend Shengyi's existing market share, but present a barrier when attempting to displace entrenched competitors.

Scale Economies yield moderate benefits by distributing fixed overhead across a multi-site facility network. However, scale alone is not a decisive differentiator in high-specification markets. While manufacturing volume provides purchasing leverage when procuring raw copper foil, high-end output is ultimately constrained by specialized raw-material availability and customer qualification approvals rather than sheer lamination capacity.

Cornered Resource represents a key structural advantage for vertically integrated competitors like Kingboard. Kingboard's internal ownership of copper foil, specialty resin, and electronic glass cloth production allows it to supply its laminate operations at cost during input supply squeezes. By contrast, Shengyi purchases raw materials at prevailing market prices. During periods when input costs inflate by over 50%, this structural cost variance generates a meaningful margin advantage for integrated producers. Furthermore, Kingboard raised roughly $122 million in July 2024 specifically to meet AI-driven circuit board demand, signaling an explicit strategy by the industry's cost leader to contest high-specification market segments.[^20]

This dynamic creates a strategic tension for Shengyi. As long as high-performance AI materials remain supply-constrained, Shengyi's process engineering and qualification credentials generate premium margins. However, if vertically integrated competitors successfully qualify comparable high-speed materials at scale, high-specification margins could compress toward commodity levels over time, shifting competitive advantage back to input cost leaders. While current market data does not indicate immediate margin compression in high-spec tiers, historical industry cycles demonstrate that material tightness eventually yields to expanding capacity.

Counter-positioning is absent in this sector, as laminate manufacturing relies on standard press architectures and business models that established competitors can readily replicate.

Branding offers modest differentiation. While hardware designers specify substrate suppliers by brand name in design documentation, customer preference is driven by technical track record and qualification history rather than consumer brand equity.

Network Economies are absent, as substrate manufacturing does not generate network effects.

In summary, Shengyi possesses one primary power in process engineering, benefits conditionally from customer switching costs, and gains incremental efficiency from manufacturing scale. However, it remains structurally vulnerable regarding cornered upstream resources during input inflation cycles. This profile characterizes a highly capable specialized manufacturer rather than an unassailable monopoly—a distinction that heavily influences valuation multiples when industry cycles turn.

Which raises the question of who is making these calls.

↑ Back to Top

VIII. Management Credibility, Capital Allocation & Governance

On June 6, 2024, Liu Shufeng retired from Shengyi Technology, citing age. He was 69. He had joined the company in January 1990 and spent thirty-four years there, arriving when annual output was measured in hundreds of thousands of square meters and leaving a group with net assets near RMB 14 billion and over ten thousand employees.¹⁰ His final full year as chairman, 2023, was a bad one — revenue of RMB 16.59 billion, down 7.93%, and net profit of RMB 1.16 billion, down 23.96%, as the electronics industry drowned in post-COVID inventory.¹⁷ He was paid RMB 8.91 million that year.¹⁰

Thirty-four-year tenures produce two kinds of successors: the outsider brought in to break the culture, or the insider raised inside it. Shengyi chose the second.

Chen Renxi

陈仁喜 Chen Renxi, born in 1967, was elected chairman and legal representative on June 7, 2024.¹⁰ His background is unusually literal for the job. He graduated from Huazhong University of Science and Technology in 1989 in organic chemistry — the actual discipline of resin formulation. He joined the Hong Kong Shengyi operation in 1995 as a market service engineer, meaning his first exposure to customers was technical support on material problems. From 1997 to 2016 he moved through production and operational roles including production director and vice president, and became chief executive in 2016, holding directorships across the group's subsidiaries in China and Thailand.¹⁰ He carries professor-level chemical engineering credentials alongside a senior economist title.¹⁰

Chen is, in other words, a resin chemist who spent twenty years running the plants that make the resin work and eight more running the company. That profile is well-matched to an industry where the competitive advantage is process knowledge, and poorly matched to a period requiring dramatic strategic reinvention. Fortunately for him, the last two years have required the former.

The transition itself has been notably uneventful, which is a compliment. There has been no strategy pivot, no restructuring charge, no unexplained divestment. The narrative running through the 2024 report, the 2025 report, and the 2026 interim disclosures is the same narrative: climb the product ladder, expand high-end capacity, keep R&D intensity constant. Consistency of narrative across cycles is one of the more reliable proxies for management credibility, and Shengyi passes it.

There is one caveat that a careful reader of the 2025 annual report should notice. Among the risks the company itself disclosed was "management structure transitions" — a phrase a company includes only when it is genuinely thinking about succession and organizational depth.¹ It is a small piece of candor and worth more than a page of confident boilerplate.

Ownership and the state question

Shengyi is often described as a state-owned enterprise. The reality is more interesting. Dongguan Guohong Investment is the controlling shareholder with 13.29%, and Hong Kong Weihua Electronics holds 12.44% — together 25.73% of a register with registered capital of RMB 2.429 billion.¹¹ Everything else is dispersed.

A controlling shareholder with 13.29% is not control in any conventional sense. It is influence over a fragmented register. The forty-year presence of the original Hong Kong partner alongside a municipal state investor produces a governance structure that is neither a classic SOE nor a founder-controlled private company — a genuine mixed-ownership arrangement of the kind Chinese reform policy has spent two decades trying to engineer deliberately and which Shengyi arrived at by historical accident in 1985.

The practical implication is a lower risk of the value-destroying behaviors associated with tightly controlled SOEs — policy-driven acquisitions, employment-preserving overcapacity, opaque related-party transactions — but a correspondingly weaker defense against a determined outside shareholder, and a board that must persuade rather than instruct. There is no disclosed evidence of related-party abuse or off-balance-sheet complexity in the group. The one structural item an activist would press on is the 62.93% stake in a separately listed downstream subsidiary, discussed earlier.

Incentives, and what is not disclosed

A word on alignment, because it is where Chinese industrial governance is most often opaque and where investors most often assume rather than verify. Shengyi discloses executive compensation in the annual report — Liu Shufeng's RMB 8.91 million for 2023 is the visible example — but the detailed performance conditions attaching to variable pay, the specific weightings between return on equity, product mix, and R&D milestones, and the vesting mechanics of any long-term incentive are not disclosed in a form that permits outside verification.¹⁰¹⁷ Investors should treat claims about precisely how management is incentivized as inference rather than fact.

What can be observed is behavior, and behavior is the better evidence anyway. Over the past decade the company has not chased volume at the expense of margin, has not diluted shareholders to fund expansion, has not issued aggressive multi-year targets it subsequently missed, and has not restated results or changed accounting policy in a way that flattered a weak year. The 2023 downturn was reported plainly, with the cause attributed to industry inventory and price competition rather than to anything more creative.¹⁷ Companies that explain bad years simply tend to be companies whose good years mean something.

The capital allocation record

Three observations, in descending order of confidence.

First, the company returns cash seriously. For 2025 it distributed approximately RMB 2.92 billion in cash dividends, equal to 87.43% of net profit.¹ That is an extraordinarily high payout for a manufacturer entering a capacity expansion cycle, and it is either impressive discipline or slightly odd timing depending on your view of the reinvestment opportunity. Given that the company is simultaneously committing RMB 5.2 billion to a single new project, the high payout implies confidence that internal cash generation and the balance sheet can carry the capex without equity issuance — a defensible position given net debt to EBITDA below 1.0x.[^1]

Second, expansion has been organic rather than acquisitive. Capital expenditure reached RMB 2.38 billion in 2025, up 156.57%, funding the Shengyi Electronics Dongcheng fifth plant and the Thailand facility among others.¹ In April 2026 the company approved its largest single investment ever: approximately RMB 5.2 billion for a high-performance CCL project at 松山湖 Songshan Lake in Dongguan, to be built in two phases, allocated roughly RMB 200 million to land, RMB 2.37 billion to production equipment, RMB 1.57 billion to construction, RMB 660 million to supporting systems, and RMB 200 million to working capital.² Management justified it by reference to an AI-grade CCL market it sized at roughly $2.2 billion in 2025 and doubling.²

Building brownfield capacity in your home city, at your technology center, in the product tier where you have qualification depth, is the lowest-risk form of expansion available. It is also the most cyclically dangerous. Every CCL producer on Earth is expanding into the same demand signal simultaneously. The industry's history is unambiguous that the capacity ordered at the top of a materials cycle arrives into the bottom of the next one. Shengyi is not making a mistake other people are avoiding. It is making the same bet everyone is making, with better-than-average odds of being on the right side of it, and investors should size the risk accordingly.

Third, there is no record of diversification into unrelated sectors — no property arm, no financial subsidiary, no strategic stake in an unrelated listed company. For a Chinese industrial group of this age and this cash generation, that restraint is genuinely uncommon and genuinely valuable.

The record, then, is of competent, conservative, technically-grounded management doing an unglamorous thing consistently for a very long time. What could go wrong is mostly not about them.

---

↑ Back to Top

IX. Material Risk Radar & Skeptical Investor Stress Test

Market rallies frequently obscure underlying operational vulnerabilities. For copper-clad laminate manufacturers, several critical operational, competitive, and structural risks warrant rigorous examination.

Raw material inflation and the integration gap. This represents the most immediate structural risk facing the business. Shengyi procures its copper foil, glass cloth, and resin systems on the open market, whereas competitor Kingboard maintains extensive vertical integration. During the input price surge in early 2026—when electronic glass cloth prices doubled, ultra-fine yarn prices doubled, high-end HVLP copper foil was allocated into 2027, and select PPE resin grades tripled—Shengyi successfully passed higher costs to customers because market demand exceeded available supply.⁷ However, cost pass-through relies on supply scarcity. Should market capacity catch up with demand, vertically integrated competitors retain a structural cost advantage that non-integrated producers cannot match. While Shengyi's margin expansion in 2025 and early 2026 demonstrated an improving product mix, its pricing power has yet to be tested in a balanced supply environment at elevated input costs.

Reference design concentration. Material selection for artificial intelligence hardware is determined well in advance of volume production, locked into reference design specifications established by chip designers and hyperscalers. EMC's dominant position across major Western AI silicon programs illustrates the strength of this specification lock-in.¹⁵ If EMC and TUC secure qualified supplier slots on upcoming architecture generations—where the friction of signal-integrity re-validation favors incumbents—Shengyi's international AI revenue growth could stall regardless of product performance. Industry reports of Shengyi achieving M9-grade qualification offer counter-evidence, though these qualifications remain unconfirmed in regulatory filings.⁵¹⁶

Geopolitical and export-control fracture. Trade friction creates dual-sided risk for the company. Tighter export restrictions on AI semiconductors flowing into China constrain the domestic server build-out that underpins the domestic substitution thesis. Conversely, tariffs or trade restrictions on Chinese-origin materials entering Western supply chains endanger the export channel that doubled in revenue in 2025.¹ While expanding manufacturing capacity in Thailand provides a partial hedge against direct export restrictions, it does not mitigate domestic silicon supply constraints. In its corporate filings, Shengyi explicitly highlighted tariff fluctuations and international regulatory compliance among its key operational risks.¹

Commodity overcapacity. Bifurcation in the copper-clad laminate market creates distinct operational exposures. While high-specification grades yield premium returns, standard FR-4 laminates remain subject to chronic industry overcapacity and price competition. Because a substantial portion of Shengyi's production volume still serves commodity applications, weakness in standard laminate pricing can weigh on consolidated earnings even during periods of strong high-end demand. The downturn in 2023—when revenue fell 7.93% and net profit dropped 23.96%—serves as a clear precedent for this cyclical vulnerability.¹⁷

Technology substitution. Advanced packaging architectures, including ceramic and glass-based substrates, present long-term substitution risks to traditional organic laminates, as acknowledged in the company's risk disclosures.¹

Receivables and working capital. Operational working capital requirements remain elevated, reflected in days of sales outstanding reaching approximately four and a half months (141.9 days) and a cash conversion cycle near 163 days.[^1] While extended payment terms function as a growth financing cost during demand expansions, elevated accounts receivable balances heighten credit and write-down risks if downstream customer demand contracts.

The stress test

A central question for investors is whether Shengyi remains fundamentally a cyclical materials producer benefiting from a temporary demand surge.

The cautious perspective rests on established historical patterns. Listed since 1998, the company has spent much of its public history generating returns on equity in the low-to-mid teens, with return on equity falling to 8.32% during the 2023 downcycle.[^1] As a non-integrated producer, it remains exposed to raw material price fluctuations in copper, resin, and glass cloth across an industry prone to periodic overcapacity. The rapid appreciation in its stock price through mid-2026 was propelled by capital expenditure cycles among hyperscale technology companies—a demand driver outside the company's direct control.[^1] Historically, cyclical material suppliers at peak demand cycles frequently attribute temporary margin expansion to permanent structural shifts.

Evaluating whether Shengyi has achieved a structural re-rating requires examining empirical operational evidence across three key areas.

First, the company demonstrates a multi-decade track record of product migration. Its transition from standard FR-4 to lead-free, halogen-free, high-glass-transition, automotive, 5G, and current M8 and M9 ultra-low-loss materials represents a sustained operational progression rather than a sudden strategic pivot. Research and development expenditure consistently averaged near 5% of revenue throughout this period, maintaining investment continuity even during industry downturns that constrained peer capital.[^1]¹

Second, financial results in 2025 reflected structural product mix expansion rather than pure volume expansion. Revenue increased 39.45% while physical laminate shipment volume grew 11.95%, demonstrating that top-line expansion was driven primarily by higher average selling prices and a greater proportion of advanced material grades.¹³

Third, margin performance during periods of input cost inflation indicates pricing power in specialized categories. While commodity manufacturers typically suffer margin compression when raw material costs rise by over 50%, Shengyi expanded consolidated gross margin by 4.52 percentage points in 2025 and by more than 14 percentage points year-over-year in the first quarter of 2026.¹⁸ This margin resilience confirms that high-specification material lines command pricing independent of base material cost-plus structures.

Ultimately, Shengyi operates as a cyclical materials enterprise with an expanding high-specification franchise embedded within its operations. Current earnings power reflects tight supply conditions in advanced substrate categories that will normalize as industry capacity expands. The central determinant of long-term value lies in the proportion of its product portfolio that achieves locked reference design status across major hardware platforms—a metric that remains protected by customer switching friction, even as overall market supply balances.

↑ Back to Top

X. Investment Playbook, Bull vs. Bear Case & Key KPIs

The bull case

The bull case rests on a single proposition: that the AI compute build-out has permanently raised the technical specifications required for circuit board materials, leaving only a small cohort of global suppliers capable of meeting those standards.

If that thesis holds, financial benefits compound across several fronts. Market share gains in ultra-low-loss material grades carry structurally higher gross margins than legacy product lines. In addition, the reported M9 qualification would position Shengyi among a select group of qualified suppliers for 224Gbps-class system architectures just as that hardware generation enters production.⁵ Under this scenario, the RMB 5.2 billion Songshan Lake facility represents timely capacity arriving to meet market tightness rather than adding to an oversupplied industry.²

The domestic market provides a second, more resilient growth vector. Chinese hyperscalers expanding AI infrastructure under trade and export restrictions require a qualified domestic substrate supplier, and Shengyi remains the primary supplier at this technical tier. This captive domestic demand provides a stable baseline that is less vulnerable to global price friction than international channels.

Third, packaging substrate development provides long-term upside potential. If Shengyi's R&D in flip-chip ball grid array (FC-BGA) materials yields commercial design wins, it would grant entry into a high-margin sector currently dominated by a Japanese and Korean oligopoly.³¹⁴ While this optionality is not reflected in current financial results and should not be priced into current equity valuation, it represents a strategic call option backed by the company's formulation capabilities.

Finally, the downstream printed circuit board subsidiary has transitioned from a capital drag into a high-margin expansion engine. In the first half of 2026, Shengyi Electronics expanded revenue by 53.46% and increased net profit by 109.36%, generating gross margins that surpassed the parent company's laminate operations.¹³¹ Having both core materials and downstream fabrication compounding simultaneously changes the group's earnings profile.

The bear case

The bear case does not depend on operational failure; it requires only that tight market conditions normalize toward historical averages.

First, the raw-material integration gap poses an ongoing margin risk. When component supply transitions from scarce to broadly available at elevated prices, competitor Kingboard's backward integration into copper foil, resin, and glass fabric yields a structural cost advantage that non-integrated producers cannot match without substantial capital and multi-year facility build-outs. Shengyi's margin expansion in 2026 reflects temporary supply tightness alongside product mix improvements, making it difficult for external investors to separate short-term scarcity rents from permanent pricing power.

Second, technical qualification does not guarantee commercial volume. Taiwan-based EMC's estimated 80% share on AWS Trainium and near-total allocation on Meta's MTIA platform stem from securing primary placement in initial reference designs.¹⁵ Displacing an incumbent requires customers to incur substantial re-validation costs and signal-integrity testing—a transition that typically occurs only during major hardware architecture refreshes every two to three years.

Third, widespread industry capacity expansion threatens future pricing discipline. High-speed laminate facilities require three to five years to develop and qualify, meaning capital projects initiated during the current cycle will come online around 2028 to 2029.⁷ Industry analysis in 2026 already indicated that supply constraints could ease by the second half of 2027.⁷ If hyperscale AI infrastructure spending decelerates before new capacity is absorbed, the industry could rapidly shift from supply allocation to aggressive price competition, eroding recent margin gains.

Fourth, current equity valuations leave limited margin of safety. After tripling from a twelve-month low of RMB 40.40 to reach RMB 143.21—and touching an intraday peak of RMB 191.88—the stock price reflects high expectations for sustained growth.[^1] Market pricing relies heavily on the 19.94% return on equity achieved in 2025 while discounting the 8.32% return on equity recorded during the 2023 cyclical downturn.[^1]

Fifth, extended working capital cycles present balance sheet exposure. With days of sales outstanding exceeding four months, any sudden deceleration in downstream customer demand could transform uncollected receivables into credit losses and inventory write-downs.

Three KPIs worth tracking

Evaluating Shengyi's multi-year trajectory relies on tracking three primary operational indicators.

1. The spread between laminate revenue growth and shipment volume growth. This metric provides the clearest empirical signal of product mix enhancement. In 2025, rigid copper-clad laminate revenue expanded 20.17% while physical sales volume grew 11.95%—generating an 8.22 percentage point positive spread.¹³ If this revenue-to-volume growth spread widens in annual disclosures, it confirms that average realization per square meter is rising through high-specification product adoption. Conversely, if the spread compresses toward zero, growth is being driven by unit volume rather than premium pricing, weakening the high-margin expansion thesis.

2. Gross margin sustainability across normalized raw-material supply. Shengyi achieved a consolidated gross margin of 25.10% in 2025 and 28.1% in the first quarter of 2026 during widespread raw-material tightness.¹⁸ The critical test will occur as upstream supply of copper foil, electronic glass cloth, and specialty resin balances. If gross margin remains in the mid-to-high twenties after supply constraints ease, it demonstrates that pricing power is structural and that the vertical integration disadvantage relative to Kingboard is manageable. If margins contract toward historical mid-teen levels, recent profitability will prove to have been driven primarily by temporary supply tightness.

3. Confirmed high-specification platform qualifications in regulatory filings. Beyond industry reports or informal investor disclosures, formal periodic filings provide the definitive record of design wins. A primary benchmark is whether future annual and interim reports confirm material qualifications in next-generation computing platforms—matching the formal disclosure standard set by subsidiary Shengyi Electronics when confirming volume production of high-order HDI and 30-plus-layer boards in its 2026 interim report.¹³ Confirmed qualification status secures multi-year demand across specific hardware architectures, whereas the absence of verified design wins in official reports would indicate that reference-design opportunities are being lost to competitors.

↑ Back to Top

XI. Epilogue & Final Lessons

Every few years, a familiar scene plays out in the Songshan Lake laboratories: an engineer retrieves a test coupon from an oven, and technicians measure how much signal survived a 30-centimeter transmission path through a composite material that did not exist five years earlier. The process is uncinematic. Yet repeating that routine across four decades is what transformed a joint venture founded among the agricultural plots of 1985 Dongguan into an industrial supplier whose market valuation approached RMB 348 billion in mid-2026.[^1]

The company's operating strategy rests on three straightforward principles: consistently migrate up the technical ladder, maintain R&D investment through cyclical downturns at roughly 5% of revenue annually,[^1]¹ and expand production capacity organically in core materials rather than pursuing conglomerate acquisitions. Component manufacturers that adhere to this framework often generate modest headlines for long periods before emerging as pivotal suppliers during major technology transitions.

Sustaining recent gains, however, presents a distinct operational challenge. Climbing the product hierarchy within a protected domestic market differs fundamentally from displacing entrenched Taiwanese competitors in the reference designs of Western AI silicon designers. Operational metrics confirm that shift has not yet occurred: in 2025, Elite Material Co. captured the top global market position while Shengyi ranked second with a 13.7% market share.⁶³ While reported M9 material qualifications indicate technical capability, commercial design wins remain to be proven in official volume disclosures.

The broader lesson recurs across hardware investment cycles. During technology expansions, market attention focuses on high-profile platforms—accelerators, foundation models, and data centers. Yet compute infrastructure remains a physical stack in which every tier can create an operational bottleneck. In 2026, one of those critical bottlenecks was a resin-impregnated sheet of glass fabric faced with ultra-thin copper foil, produced by a small cohort of qualified vendors and priced by supply constraints. Substrates that enable high-speed signal integrity warrant close monitoring alongside the silicon they support—at least until industry capacity expands to meet demand.

---

References

1. 生益科技2025年报解读：营收增39.45%净利润翻倍 现金流净额大增262.17% — 新浪财经, 2026-04-25 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

2. 生益科技2025年净利润增长超九成 52亿元投资加码高性能覆铜板布局 — 证券时报, 2026-04-24 ↩↩↩↩↩

3. 【企业动态】生益科技2025年净利润增长超九成 52亿元投资加码高性能覆铜板布局 — 东方财富网, 2026-04-24 ↩↩↩↩↩↩↩↩↩↩↩

4. 生益科技：上半年净利润同比预增117%到131% — 证券时报, 2026-07-13 ↩

5. 英伟达"弃用M9"传言背后，覆铜板产业链正迎来一场AI算力材料革命 — OFweek电子工程网, 2026-06 ↩↩↩↩↩↩↩↩

6. EMC's Rise to No. 1: AI Servers Redefine the Global CCL Supply Chain — TSPA Semiconductor, 2026 ↩↩↩↩↩↩↩

7. 覆铜板涨价潮：建滔、生益订单排满，上游原材料全线紧缺 — 新浪新闻, 2026 ↩↩↩↩↩↩↩↩↩↩↩

8. 生益科技（600183）2026年一季报简析：营收净利润同比双双增长，盈利能力上升 — 腾讯新闻, 2026-04-30 ↩↩↩↩↩

9. Shengyi Technology Official Corporate Portal — SYTECH, 2026 ↩↩

10. 刘述峰退休离任，陈仁喜接棒生益科技董事长 — 南方都市报, 2024-06 ↩↩↩↩↩↩↩↩

11. 广东生益科技修订公司章程 注册资本24.29亿元 控股股东持股13.29% — 新浪财经, 2025-10-28 ↩↩

12. Shengyi Electronics Co., Ltd. STAR Market IPO Prospectus & Disclosure — Shanghai Stock Exchange, 2021-02-02 ↩↩

13. 生益电子上半年归母净利润同比增长109.36% AI PCB为增长主线 — 新浪财经, 2026-08-14 ↩↩↩↩↩

14. CCL Market Update October 2025: 2024 Market Rebound Driven by High-Speed and AI Demand — Prismark Partners LLC, 2025-10 ↩↩↩↩

15. Elite Material, TUC race ahead in AI CCL push — DIGITIMES, 2026-07-20 ↩↩↩↩

16. M10覆铜板是否送样英伟达？生益科技回应 — 新浪财经, 2026-07-06 ↩↩↩↩

17. Shengyi Technology Investor Relations & Official Disclosure Portal — Cninfo Disclosure, 2026 ↩↩↩↩↩↩