China Resources Microelectronics: The IDM Titan of Chinese Silicon

I. Introduction & Episode Roadmap

In the summer of 1997, in a low-slung industrial compound on the western edge of Wuxi, Jiangsu province, engineers finally switched on a six-inch wafer line that the Chinese state had spent seven years and roughly RMB 2 billion attempting to build. It was intended to mark China's transition into global semiconductor manufacturing. Instead, it demonstrated the relentless pace of semiconductor development cycles. By the time the tools were operational, their underlying process technology lagged four to five generations behind what Japanese, American, and European commercial fabs were shipping. The line produced just a few hundred wafers per month against a monthly design target of 12,000, losing money from day one.

That compound became 无锡华晶 Wuxi Huajing. Wuxi Huajing evolved into 华润上华 CSMC Technologies. CSMC eventually became 华润微电子有限公司 China Resources Microelectronics Limited — CR Micro — which today trades on the 上海证券交易所科创板 Shanghai Stock Exchange STAR Market under ticker 688396. Market research firm Omdia ranks CR Micro as the largest MOSFET manufacturer in China and the country's second-largest power semiconductor producer. In the first half of 2026, the company reported revenue of RMB 6.13 billion.¹²

That is the corporate arc. It is a compelling narrative, but for an investor, it is only the starting point.

The central mystery lies in the business model. For twenty-five years, the global semiconductor industry has reorganized around a single principle: separating chip design from manufacturing. 台积电 TSMC demonstrated that a pure-play foundry could out-invest and out-yield integrated rivals; 中芯国际 SMIC replicated that model in China; and hundreds of fabless design firms prospered without owning cleanrooms. CR Micro defied that consensus by doubling down on the Integrated Device Manufacturer (IDM) model — 设计制造封测一体化, integrating design, fabrication, packaging, and testing under one roof — specifically for power semiconductors. This approach represents the most capital-intensive way to compete in one of the semiconductor market's most price-sensitive segments.

Why did CR Micro take this path, and has it proven successful?

A second mystery surrounds its capital markets history. CR Micro is not a newcomer to public exchanges. It listed in Hong Kong in 2004, spent seven years largely overlooked by investors, and was taken private by its state-owned parent in 2011. Nine years later, it returned to public markets on the Chinese mainland as the first offshore-incorporated 红筹企业 red-chip enterprise admitted to the STAR Market.³ The same underlying assets, under a new trading venue and policy environment, achieved valuations at a multiple of what Hong Kong investors had previously paid. Whether that repricing reflected a structural transformation of the business or policy-driven valuation expansion remains contested after six years of trading. By late August 2026, the stock traded near RMB 59 per share—giving the company a market capitalization of roughly RMB 79 billion—after fluctuating between RMB 43.71 and RMB 101.99 over the preceding twelve months. That wide trading band highlights how undecided the market remains on the company's true value.⁴

This analysis traces the company's complete trajectory: from its state-backed origins and the Huajing rescue to its strategic pivot toward power semiconductors under China Resources ownership; from its red-chip listing and pandemic supercycle to its large, structurally complex twelve-inch fab expansions in Chongqing and Shenzhen; from optionality bets in silicon carbide and smart control to a competitive assessment using Porter's Five Forces, Hamilton Helmer's 7 Powers, and a stress test of what skeptical investors view as the state-owned enterprise CapEx trap.

Two structural features warrant emphasis at the outset, as they underpin much of the subsequent analysis.

The first is the dual-segment business structure. CR Micro sells its own branded chips through its 产品与方案 Products and Solutions division, while simultaneously offering foundry, packaging, and photomask services to external semiconductor companies through its 制造与服务 Manufacturing and Services division.¹ In practice, it operates as both a device manufacturer and a contract foundry sharing the same cleanroom infrastructure. Which of these two segments generates superior profitability provides key insights into the company's true economics—and contradicts common market assumptions.

The second feature involves ownership structure. The two twelve-inch fabs central to CR Micro's growth strategy are not majority-owned subsidiaries. Instead, they are minority equity stakes in separately capitalized joint ventures funded primarily by municipal state entities. Consequently, they appear on the company's financial statements as a share of joint-venture losses rather than consolidated plant assets.¹ This accounting treatment significantly influences the balance sheet, income statement, and overall investment thesis.

Throughout this evaluation, one standard applies: every management claim regarding the company's competitive moat is evaluated against its disclosed execution record. CR Micro possesses a long operational history, much of which points to important caveats.

Subscribe on LinkedIn

↑ Back to Top

II. State Roots & The Heritage of Wuxi Huajing (1980s – 2001)

Understanding why CR Micro exists requires examining what Chinese state planners sought to address at the dawn of the country's modern industrial push.

In 1990, the Ministry of Electronics Industry launched 九零八工程 Project 908 as the semiconductor cornerstone of the Eighth Five-Year Plan. The state committed roughly RMB 2 billion—about $418 million at the time—to Huajing, the operating entity of Wuxi Factory No. 742. The mandate was to construct a six-inch wafer line capable of producing 12,000 wafers per month at 0.8 to 1.2 micron geometries, relying on process technology transferred from AT&T and later Lucent.⁵ On paper, the strategy appeared coherent: purchase a proven manufacturing process, train domestic engineers, and establish a foundational national fab.

Instead, Project 908 fell victim to prolonged administrative execution. Planning approvals, equipment orders, and physical construction dragged on for roughly eight years—a devastating delay in an industry where product generations cycle in less than two.⁵ When the line finally went live in 1997, the commercial market had already advanced. Production volume reached only a fraction of its design target, leaving the fab to incur immediate operational losses.

The underlying economic failure stemmed from the rigid capital dynamics of semiconductor manufacturing. Wafer fabs operate as high fixed-cost assets. Capital expenditure is committed up front and depreciated over five to ten years regardless of output volume, meaning unit economics depend almost entirely on shipping sufficient high-yield wafers to cover that fixed base. Huajing carried the heavy depreciation overhead of a commercial fab alongside the low throughput of a pilot plant—a structural mismatch no operational cost-cutting could resolve.

Beyond utilization challenges, Huajing lacked a commercially defensible product line. Although its technical teams made genuine strides in bipolar processing and basic device physics—while training a generation of process engineers who later built the Jiangsu semiconductor cluster—technical capability alone could not offset ongoing depreciation.

The competitive backdrop deteriorated further as China prepared for World Trade Organization accession. Lower import tariffs exposed legacy state-owned fabs to direct competition from established global power semiconductor manufacturers, including 英飞凌 Infineon Technologies, 意法半导体 STMicroelectronics, and 安森美 ON Semiconductor. These global competitors possessed decades of refined process recipes, fully depreciated manufacturing equipment, and extensive distribution networks capable of delivering qualified components on demand. For Chinese appliance makers in 2001, selecting a MOSFET was a commercial decision based on price and reliability data rather than policy preference, and imported components consistently won.

By the late 1990s, Huajing's financial position was unsustainable. Resolution arrived through private capital and external operational expertise. In 1999, Dr. 陈正宇 Chen Zhengyu partnered with Huajing to form 无锡华晶上华半导体 Wuxi Huajing CSMC Semiconductor, a joint venture designed to operate a six-inch MOS foundry on Huajing's existing production line.⁶ The strategic shift was straightforward: rather than attempting to fill cleanroom capacity with self-designed products, the company rented its manufacturing line to third-party chip designers.

That pivot from an underperforming product manufacturer to a contract foundry established the structural foundation for CR Micro's current business. Today, the company maintains this dual operational model, dividing its business between internal device production in its 产品与方案 Products and Solutions division and contract manufacturing in its 制造与服务 Manufacturing and Services division.¹

The trajectory of Project 908 underscores a key lesson for semiconductor investors: capital deployed on slow approval cycles destroys value, as assets depreciate against a fast-moving technology clock. This dynamic remains directly relevant to the multi-billion-dollar twelve-inch fab expansions CR Micro is financing three decades later.

By 2001, Huajing required an owner equipped with capital, patience, and commercial discipline. It found that partner in China Resources, a Hong Kong-based conglomerate then best known for consumer retail, beer, and real estate.

↑ Back to Top

III. The China Resources Rescue, CSMC Consolidation, & The HKEX Odyssey (2002 – 2011)

China Resources Group (华润集团) occupies a distinct position within the Chinese state sector. Founded in Hong Kong as a state trading enterprise, it evolved into a central state-owned enterprise supervised by the State-owned Assets Supervision and Administration Commission (国资委 SASAC), building a portfolio spanning power generation, property, cement, gas, pharmaceuticals, beer, and retail. It operated with the commercial reflexes of a Hong Kong conglomerate rather than the bureaucratic pacing of a mainland ministry. That combination—a state balance sheet, commercial operating discipline, and offshore incorporation—offered precisely what the distressed Wuxi fab required.

In 2002, China Resources indirectly acquired the entirety of Huajing, taking over the Project 908 assets and operating them alongside Chen Zhengyu's foundry venture.⁶ The following year, the group restructured the mainland entities into an offshore holding company, CSMC Technologies, explicitly to create a listable vehicle. In August 2004, CSMC completed its initial public offering on the main board of the Hong Kong Stock Exchange, issuing 684 million shares and raising approximately HK$342 million.⁶

A capital raise of HK$342 million—roughly $44 million at the time—was strikingly modest for a semiconductor manufacturer. For an enterprise seeking to industrialize Chinese wafer fabrication, $44 million represented less than the cost of a few specialized manufacturing tools rather than transformative growth capital. The small offering provided the first piece of evidence contradicting the narrative that the listing solved the company's financing needs.

Testing the claim: did the Hong Kong listing validate the strategy?

The bullish reading of this era—often highlighted in sell-side histories—argues that the Hong Kong listing validated CR Micro's open-foundry model, opened access to international capital, and financed the six-inch and eight-inch wafer line upgrades that established the modern enterprise. The historical record demonstrates otherwise.

Beyond the limited capital raised, the operational financial performance across subsequent years underscores the ongoing challenge. When CR Micro eventually filed to return to mainland public markets, its prospectus detailed the financial condition resulting from its Hong Kong era: by the end of 2018—seven years after its Hong Kong delisting and fourteen years after its initial public offering—the consolidated balance sheet carried accumulated losses of RMB 3.05 billion.⁶ The company incurred net losses in both 2016 and 2017, turning its first meaningful net profit only in 2018.⁶ Furthermore, two of its primary operating subsidiaries—the Wuxi CSMC foundry and the Chongqing power device entity—carried accumulated losses of RMB 2.73 billion and RMB 2.12 billion, respectively, which under Chinese corporate law barred them from distributing cash dividends upward until those deficits were cleared.⁶

Fifteen years of capital-intensive competition in commodity power semiconductors had not produced an economic moat; it had generated substantial balance-sheet deficits.

The limited profitability achieved during this window was heavily reliant on non-operating support. From 2016 through 2018, CR Micro recognized RMB 266 million in government subsidies, compared to a cumulative net profit of RMB 132 million over those same three years.⁶ Excluding government grants, the underlying business remained loss-making in the aggregate. While government subsidies are standard and fully disclosed across the domestic semiconductor sector, this reliance highlights the fragile quality of the earnings base prior to its subsequent market re-rating.

Public markets reflected these structural challenges. Trading under ticker 00597, the stock remained illiquid and depressed during the post-2008 industry downturn.⁸ China Resources attempted a privatization in 2009 that failed to secure shareholder approval. The parent group returned in 2011 with an increased offer of at least HK$0.48 per share—compared to approximately HK$0.30 in the earlier attempt—and completed the privatization and delisting from the Hong Kong Stock Exchange main board in November 2011.⁶⁸ The aggregate transaction value was not specified in public disclosures from the period.

That decade highlights a broader dynamic common to Chinese industrial listings of the era. A public listing provides access to capital, but that access generates value only when public investors are willing to fund the business. In the late 2000s, Hong Kong institutional investors prioritized high-growth Chinese consumer brands and property developers. They demonstrated little appetite for a sub-scale, capital-intensive, loss-making wafer fabricator with state ownership and an unproven market position. Trapped in a self-reinforcing cycle—where a depressed valuation prevented meaningful equity raises, and a lack of capital constrained operational improvements—privatization offered a logical exit for the parent group.

The calibrated conclusion. The claim that the Hong Kong listing validated the business strategy is contradicted by the financial record. The listing's actual contribution was operational and structural: it consolidated fragmented state assets into an offshore-incorporated, audited corporate structure and subjected the enterprise to seven years of public market reporting discipline. That entity—a Cayman Islands-incorporated holding company owning mainland operating subsidiaries—established the exact red-chip framework that enabled CR Micro's subsequent return to mainland domestic exchanges. The Hong Kong chapter did not capitalize the company's growth; it established its corporate foundation.

Which raises the central question: if the strategy through 2011 failed to deliver commercial profitability, what changed?

↑ Back to Top

IV. Strategic Pivot: Building China's #1 Power Semiconductor IDM (2011 – 2019)

Between its Hong Kong delisting and its eventual mainland return, CR Micro shifted away from imitating advanced logic foundries.

The critical choice was defining what not to pursue. During the 2010s, China's national semiconductor policy focused heavily on advanced digital logic—nanometer scaling, extreme ultraviolet (EUV) lithography, and closing the gap with TSMC. Lacking the capital, advanced tooling access, and customer base required for leading-edge nodes, CR Micro focused on its existing assets: legacy six-inch and eight-inch wafer lines, a device-design heritage stemming from Wuxi Factory No. 742, and packaging operations. It pivoted toward segments less bound by Moore's Law: power semiconductors and specialty analog chips.

Why the physics changes the business model. The economic rationale for this pivot rests on device physics. In digital logic, performance advances primarily by shrinking features—packing more transistors onto a die to increase processing speed and lower power consumption. Because logic scaling is essentially a lithography challenge requiring massive capital expenditure, manufacturing consolidated into a few well-capitalized foundries. Design and manufacturing split cleanly because their interface is standardized: a fabless design firm transfers a digital layout to a merchant foundry, which executes a standard process.

Power devices operate under different physical constraints. Components such as MOSFETs and insulated-gate bipolar transistors (IGBTs) act as high-power switches, rapidly turning heavy electrical currents on and off without overheating. Performance depends less on minimum feature size and more on managing heat and voltage through three-dimensional device design: trench depth, silicon doping profiles, and thermal packaging. Consequently, device design and process engineering are deeply interdependent.

This interdependence blurs the boundary between design and manufacturing. A competitive power device cannot be specified solely through a standardized design layout handed to an external fab, as key performance metrics depend directly on proprietary fab process parameters.

To draw an analogy, designing a logic chip resembles writing a detailed recipe for any equipped kitchen to execute. Designing a power device is closer to winemaking: grape selection matters, but so do soil composition, barrel aging, and cellaring. The process and product are too tightly linked to easily decouple.

This interdependency forms the core argument for the IDM model in power semiconductors, as CR Micro notes in its disclosures: internal integration accelerates development cycles, eliminates redundant silicon-validation cycles, and allows process recipes to be tailored to specific customer specifications.¹

While logically sound, this IDM advantage exhibits a clear structural limit.

The integrated model delivers clear value at the performance frontier—such as super-junction MOSFETs, shielded-gate trench structures, and high-voltage IGBT platforms—where minimizing on-resistance and switching losses requires tight co-optimization between design and process.

In commodity, mid-tier market segments, however, fabless designers using mature foundry processes can produce adequate components with lower fixed overhead, frequently undercutting IDMs on price. As a result, the practical advantage of the IDM model is more narrow than broad corporate marketing suggests.

What the pivot actually delivered. By 2019, prior to its mainland listing, CR Micro reported annual revenue of RMB 5.74 billion, net profit attributable to shareholders of RMB 401 million, and a weighted average return on equity of 8.03 percent.⁹ Generating a 7 percent net margin on a capital-intensive asset base reflected modest corporate profitability rather than dominant pricing power. Nevertheless, these figures marked a clear operational turnaround from the net losses recorded in 2016 and 2017.

The company's sales volume expanded across broad industrial and consumer categories. Power devices serve virtually any system managing electrical current, including home appliances, power supplies, industrial motor drives, lighting, welding equipment, uninterruptible power supplies, electric two-wheelers, and increasingly, solar inverters and electric vehicles.¹ These end markets offer massive unit volume alongside price sensitivity, where customer purchasing decisions depend on qualification history, lead times, and unit costs. Through steady capacity scaling, CR Micro expanded its market presence: an October 2025 ranking by research firm Omdia placed CR Micro first among domestic MOSFET suppliers by sales volume and second among Chinese power semiconductor producers overall.¹

The segment structure, and why it matters more than it looks. The dual-segment structure inherited from the Wuxi Huajing restructuring evolved into a core corporate strategy. Under this structure, the Products and Solutions segment markets CR Micro's proprietary devices, while the Manufacturing and Services segment sells excess fab capacity to external customers through contract foundry work, packaging and testing, and photomask manufacturing.¹ Management's rationale holds that contract manufacturing absorbs excess cleanroom capacity during cyclical downturns in proprietary device demand, stabilizing overall fab utilization.

However, historical disclosures reveal important margin dynamics between these units. CR Micro's 2019 prospectus showed that gross margins in its foundry and services segment lagged peer averages, rising from 9.9 percent in 2016 to 18.6 percent in 2018, compared to industry peer averages of 22.2 percent and 33.5 percent over the same period.⁶ In contrast, the Products and Solutions segment generated a 34.0 percent gross margin in 2018, outpacing the 29.4 percent peer average.⁶ During this phase, contract manufacturing served primarily to absorb fixed overhead, while branded device sales generated corporate profitability—a margin relationship that subsequently shifted.

Channel distribution dynamics further clarify the durability of the company's market position. Power semiconductors in China ship through two main routes: direct sales to large industrial and appliance manufacturers requiring extensive product qualification, and indirect sales through component distributors targeting smaller buyers. CR Micro established a strong position in direct accounts, particularly with intelligent power modules (IPMs)—integrated switch-and-driver circuits used in variable-frequency air conditioner and refrigerator compressors. By 2025, the company reported that appliance and industrial control applications accounted for over 80 percent of its IPM sales, securing placement in major product lines across multiple tier-one appliance brands.¹ While these direct relationships offer customer switching costs, they represent a diminishing share of sales within a broader product mix increasingly exposed to competitive distribution channels.

Expanding beyond these legacy lines required growth capital far exceeding what the company had raised in Hong Kong. In 2019, mainland financial regulators launched a new equity venue designed specifically for domestic technology enterprises.

↑ Back to Top

V. The Landmark Red-Chip STAR Market IPO & The Supercycle (2020 – 2022)

The STAR Market opened in July 2019 as Beijing's answer to a persistent challenge: China's leading technology companies routinely sought overseas listings. For historical reasons linked to foreign venture capital financing, many operated under offshore structures—the classic red-chip framework featuring a Cayman Islands or British Virgin Islands holding company positioned above mainland operating subsidiaries. Existing regulations effectively barred those vehicles from listing on domestic A-share exchanges. The STAR Market's revised framework removed that barrier, establishing the path for an initial issuer.

CR Micro led that transition. On February 27, 2020, during the initial COVID-19 lockdown while the exchange conducted the ceremony virtually, CR Micro became the first red-chip enterprise admitted to the Shanghai Stock Exchange.³ The initial public offering priced at RMB 12.80 per share for 292.99 million shares—expanding to 336.94 million shares upon full exercise of the over-allotment option—valuing the company at roughly RMB 15 billion and implying a price-to-earnings multiple of 48.3 times trailing 2018 earnings.¹⁰ Exercising the over-allotment option brought gross proceeds above RMB 4.3 billion.

Public trading responded rapidly to the policy-backed offering. The stock closed its first trading session at RMB 42.01 per share, representing a 228 percent single-day gain.¹¹ Within a single trading day, an enterprise that public markets in Hong Kong had valued in the hundreds of millions of Hong Kong dollars reached a market capitalization of approximately RMB 49 billion.

On the arbitrage. Viewing this repricing strictly as a masterstroke of capital markets engineering—buying an underappreciated asset off an illiquid offshore exchange, holding it privately for nine years, and listing it at a premium on a policy-supported domestic board—captures the financial mechanics, but oversimplifies the underlying drivers. Between 2011 and 2020, three distinct variables shifted simultaneously: operational performance improved, the trading venue changed, and domestic semiconductor policy intensified dramatically. Management controlled only the operational improvement. The corporate re-rating stemmed primarily from where and when the equity was offered rather than a steady compounding of intrinsic value. That distinction does not diminish the execution of the listing, but it cautions against viewing the valuation step-up as a repeatable strategy.

The allocation of proceeds also merits examination. The initial 2019 prospectus targeted RMB 3.0 billion in gross proceeds: RMB 1.5 billion for an eight-inch high-end sensor and power semiconductor project, RMB 600 million for technology and product R&D, RMB 300 million for acquisitions, and RMB 600 million for working capital.⁶ The initial scope remained focused on expanding eight-inch capacity rather than twelve-inch manufacturing. The capital-intensive twelve-inch wafer program that subsequently came to define CR Micro's balance sheet was not part of the initial public offering mandate.

The supercycle. The market timing of the listing proved exceptionally favorable over the short term. Global semiconductor shortages, pandemic-driven demand for consumer electronics, accelerating electric vehicle adoption, and expanding solar installations converged simultaneously. In 2021, revenue grew 32.6 percent to RMB 9.25 billion, net profit attributable to shareholders more than doubled to RMB 2.27 billion, and blended gross margin expanded by 7.86 percentage points in a single year on high capacity utilization, an improved product mix, and higher average selling prices.⁹ Return on equity reached 15.5 percent—the highest of the company's modern era—while operating cash flow reached RMB 3.45 billion.⁹ The expansion peaked in 2022, with revenue reaching RMB 10.06 billion, net profit climbing to RMB 2.62 billion, and return on equity standing at 14.0 percent.¹²

For two years, the company's financial performance matched the bullish expectations surrounding the listing.

Understanding why the upturn proved so profitable explains why the subsequent contraction was equally sharp. During a supply shortage, a power semiconductor manufacturer benefits across three operational levers at once. High fab utilization reduces fixed depreciation costs per wafer; supply constraints allow producers to raise prices; and limited capacity enables fabs to prioritize high-margin products over lower-value lines. These factors compound simultaneously. That compounding leverage explains how a 33 percent revenue expansion in 2021 translated into a 135 percent surge in net profit.⁹

When market conditions soften, all three levers reverse together.

Testing the claim: does the IDM model insulate profitability through the cycle?

This assertion forms a central pillar of the investment thesis, and subsequent financial results provide a direct test of its validity.

In 2023, revenue contracted 1.6 percent to RMB 9.90 billion, while net profit dropped 43 percent to RMB 1.48 billion, driving return on equity down to 7.15 percent.¹² In 2024, revenue edged up to RMB 10.12 billion, but net profit fell an additional 48 percent to RMB 762 million, pulling return on equity down to 3.47 percent as blended gross margin contracted 4.76 percentage points to 27.49 percent.¹² By 2025, revenue grew 9.2 percent to a record RMB 11.05 billion, yet net profit fell another 13 percent to RMB 661 million—with non-recurring items excluded, profit dropped 27 percent to RMB 471 million—as gross margin slipped to 26.51 percent and return on equity fell to 2.91 percent.¹

The divergence between top-line sales and bottom-line profit highlights the shift. Revenue in 2025 stood 19 percent above its 2022 peak, whereas net profit was 75 percent lower.

This outcome demonstrates fixed operating leverage working in reverse, contradicting claims that the integrated model insulates earnings from industry downturns. Rather than dampening cyclical swings, the IDM structure amplifies them in both expansionary and contractionary phases.

When end-market demand weakens and domestic competitors add mature-node capacity, cleanrooms must still be operated, equipment continues to depreciate, and engineering teams remain on payroll. By contrast, a fabless designer facing a downturn can reduce foundry orders, aligning its cost structure with lower sales within a single quarter.

An integrated device manufacturer cannot adjust its fixed cost base as rapidly. That outcome reflects the financial physics of owning manufacturing assets rather than an operational misstep.

Assertions that gross margins collapsed toward 20 percent overstate the contraction; the reported blended gross margin reached its trough at 26.5 percent in 2025, compared to a 2021 peak that management disclosed as a 7.86 percentage point expansion year over year.¹⁹ However, net earnings suffered a disproportionate decline because below the gross margin line lay growing depreciation expenses and equity-method losses from non-consolidated joint ventures—most notably the heavy capital commitments in Chongqing and Shenzhen.

↑ Back to Top

VI. Mega-CapEx: Chongqing & Shenzhen 12-Inch Fabs & Management

A central feature of CR Micro’s balance sheet structure is that its flagship twelve-inch wafer fabrication facilities are not consolidated assets. Instead, they are held as equity-method investments, contributing a proportional share of net profits or losses to the income statement without adding the facilities' underlying assets, liabilities, or revenues to the consolidated financial statements. CR Micro holds a 19 percent equity stake in 润西微电子（重庆）Runxi Microelectronics (Chongqing), which operates the twelve-inch power device fabrication line, and a 33 percent stake in 润鹏半导体（深圳）Runpeng Semiconductor (Shenzhen), which operates the twelve-inch specialty analog line.¹ The registered capital of the Chongqing and Shenzhen corporate entities stands at RMB 5 billion and RMB 15 billion, respectively.¹ The remaining equity in both projects is held by local government funds, state industrial investment funds, and affiliated China Resources entities.

This structure presents a dual financial dynamic. On one hand, CR Micro secures access to advanced twelve-inch manufacturing capacity—gaining scale advantages that yield roughly 2.25 times more usable silicon area per wafer than legacy eight-inch lines—without carrying RMB 20 billion in plant construction costs or project debt directly on its balance sheet. Consequently, reported net debt remains conservative, while municipal investment entities in Chongqing and Shenzhen absorb the bulk of the initial capital risk. CR Micro subsequently procures finished wafers from these joint ventures as a corporate customer, purchasing RMB 688 million in goods from the Chongqing facility in 2025, up from RMB 404 million in 2024, and RMB 57 million from the Shenzhen facility.¹

On the other hand, the arrangement does not eliminate underlying manufacturing economics; it shifts their financial visibility below the operating line. Because CR Micro serves simultaneously as a 19 percent equity holder in the Chongqing fab and as one of its primary commercial customers, wafer purchases are executed at negotiated prices.¹ While these transactions are structured as arms-length commercial arrangements and disclosed under related-party reporting guidelines, transfer pricing between the entities directly influences where corporate profits are recognized—a governance structure in which public minority shareholders hold differing equity stakes in the parent company versus the underlying project vehicles.

The near-term financial impact of this twelve-inch expansion on shareholders has been substantial. In 2025, the Chongqing joint venture generated RMB 701 million in revenue alongside a net loss of RMB 603 million, while the Shenzhen facility recorded RMB 61 million in revenue against a net loss of RMB 1.20 billion.¹ CR Micro's recognized share of losses from equity-method investees reached RMB 503 million in 2025, expanding from RMB 390 million in 2024.¹ Compared to the company's reported net profit attributable to shareholders of RMB 661 million in 2025, equity-accounting losses from the two projects absorbed three-quarters of net earnings before factoring in consolidated depreciation across internal fabs, packaging lines, and photomask operations.

This profit drag persisted into 2026. During the first half of 2026—a period marked by broader sector demand recovery, two rounds of product price increases, and high plant utilization—equity-method losses expanded to RMB 265 million from RMB 232 million in the prior-year period, driven by a RMB 339 million net loss at Chongqing and a RMB 646 million net loss at Shenzhen.² Although Shenzhen's first-half revenue of RMB 167 million nearly tripled its full-year 2025 total, top-line growth remained insufficient to offset the depreciation overhead of a RMB 15 billion capital base.

Operationally, the facilities continue to scale. By mid-2026, the Chongqing twelve-inch facility reached full production capacity of approximately 30,000 wafers per month, while the Shenzhen facility accelerated its volume ramp toward a target of 40,000 wafers per month.¹³ Management set a target in the 2025 annual report to exceed 30,000 wafers per month at Shenzhen by the end of 2026 while advancing process engineering for next-generation technology nodes.¹ Concurrently, CR Micro maintained consolidated internal capacity of roughly 230,000 six-inch wafers and 140,000 eight-inch wafers per month, while progressively reallocating six-inch cleanroom lines to silicon carbide production.¹¹³

Technologically, the two lines target distinct product categories. The Shenzhen facility focuses on specialty analog platforms at process nodes of 55 to 40 nanometers and above—including Bipolar-CMOS-DMOS (BCD), high-voltage, radio frequency, and embedded flash processes tailored for microcontrollers, power management integrated circuits, gate drivers, and communication components.¹ The Chongqing facility prioritizes higher-value silicon power devices, manufacturing super-junction MOSFETs, shielded-gate trench MOSFETs, and insulated-gate bipolar transistors (IGBTs) designed for renewable energy systems, industrial automation, AI computing infrastructure, and automotive power electronics.¹

The assertion that these multi-billion-renminbi twelve-inch investments deliver immediate accretive returns for equity shareholders is not supported by current financial results. Returns on capital have compressed under the combined weight of fixed asset depreciation and capacity ramp costs: weighted average return on equity declined from 15.5 percent in 2021 ⁹ to 2.91 percent in 2025 ¹, standing at 3.11 percent in the first half of 2026, with core operations excluding non-recurring items contributing 1.78 percent.² Management explicitly acknowledged this pressure in its 2025 annual report, noting that heavy fixed-asset depreciation had exerted a clear drag on overall profitability metrics.¹

The long-term investment case depends on whether volume scaling can convert this fixed cost base into operating leverage. If the Chongqing plant sustains full capacity utilization and Shenzhen exceeds 30,000 wafers per month with higher proportions of automotive and industrial components, equity-method investment drag would narrow and eventually turn accretive. Progress toward this objective will be visible in the half-yearly net performance disclosures of the two associate ventures.

These capital expansion initiatives coincided with major executive leadership changes in early 2025. On March 28, 2025, the board elected He Xiaolong as chairman, succeeding Chen Xiaojun following a corporate reassignment.¹⁴ He Xiaolong holds a master's degree in microelectronics from the University of Electronic Science and Technology of China and carries senior engineering credentials; his career includes policy roles overseeing semiconductor materials and chip design within China's Ministry of Industry and Information Technology, as well as executive strategy roles within China Resources.¹⁴

Six days later, on April 3, 2025, president Li Hong resigned from his positions as director and president, citing personal reasons.¹⁵ A US citizen with a doctorate in physics from the University of Wisconsin and nearly thirty years of industry experience, Li joined CR Micro in 2009 and served as president starting in March 2022.¹⁵ The company stated that his departure would not disrupt ongoing operations, as his responsibilities centered on executive management rather than primary technology development.¹⁵

Governance adjustments continued into 2026. On August 25, 2026, coinciding with the publication of interim financial results, the company announced the resignation of independent director Li Yue for personal reasons; Li chaired the audit and compliance committee while serving on the strategy, sustainability, and remuneration committees.¹⁶ The company confirmed that no unfulfilled public commitments remained outstanding upon Li Yue's departure.¹⁶

At the end of 2025, China Resources (Microelectronics) Limited retained controlling ownership with a 66.21 percent equity stake. The National Integrated Circuit Industry Investment Fund held 1.00 percent, while the Chongqing Xiyong Microelectronics Industrial Park development entity held 0.90 percent, with 12 million shares pledged.¹ Ultimate governance rests with China Resources under the supervision of SASAC. The company's external auditor, BDO China, issued an unqualified audit opinion on the 2025 financial statements.¹

Capital return policies have maintained modest continuity through the earnings contraction. CR Micro distributed cash dividends totaling RMB 102 million for 2025, representing 15.5 percent of net profit, and declared an interim dividend of RMB 0.056 per share for the first half of 2026.¹¹⁷ In addition, in May 2026, the company committed RMB 194 million—representing 16.16 percent of total fund capital—to a new Chongqing industrial fund alongside municipal investment vehicles, a central state-owned enterprise fund, and an investment manager affiliated with China Resources, expanding its strategic commitments within the regional semiconductor supply chain.¹⁸

↑ Back to Top

VII. Optionality Bets: Third-Generation Semiconductors & Smart Controls

Most power semiconductor manufacturers present a similar narrative regarding silicon carbide: traditional silicon faces physical performance limits at higher voltages, whereas wide-bandgap materials like silicon carbide (碳化硅) and gallium nitride (氮化镓) switch faster, withstand higher temperatures, and reduce power loss. Because electric vehicles, solar inverters, and artificial intelligence data centers demand those exact characteristics, management teams argue that their silicon carbide production lines serve as a primary corporate growth engine.

Every premise in that narrative is technically accurate—except the conclusion that it translates directly into superior profitability.

The physics, briefly. Conventional silicon power switches face intrinsic material limits on energy efficiency. Silicon carbide features a wider bandgap—essentially a higher electrical breakdown strength—allowing a device to withstand high voltages with a much thinner, less resistive layer of material. In practice, this design reduces heat dissipation, shrinks cooling requirements, enables higher switching frequencies, and permits smaller magnetic components. In electric vehicles, replacing conventional silicon switches with silicon carbide in traction inverters expands driving range without increasing battery size. In AI server power supplies, where energy conversion efficiency must exceed 97 percent, silicon carbide and gallium nitride are becoming essential to meet design standards.¹

CR Micro's technological execution is tangible. By the end of 2025, the company had commercialized silicon carbide MOSFET platforms rated at 650, 750, and 1,200 volts, while placing 1,700-volt and 2,200-volt platforms into production for data centers and energy storage systems. Concurrently, management reported that annual revenue from silicon carbide devices and modules more than doubled year on year.¹ The company shipped silicon carbide Schottky barrier diodes into solar inverter, storage, and EV charging infrastructure, while slating fifth-generation silicon carbide MOSFETs and trench JFETs for volume production in 2026.¹ On gallium nitride platforms, CR Micro brought an epitaxy center online, scaled a 650-volt depletion-mode platform to commercial production, and introduced a fourth-generation product family targeted at AI servers, on-board chargers, lidar, and robotics.¹ During the first half of 2026, automotive and energy-storage sectors generated 85 percent of silicon carbide revenues, supported by a fast-charging design win and volume deliveries of fourth-generation components to EV charging-station customers.²

This technical record is solid. The critical step is testing whether these technical advances translate into commercial profitability.

Testing the claim: is SiC an immediate high-margin growth engine?

The financial metrics present a clear contrast. The Products and Solutions division—which includes all self-branded silicon MOSFETs, IGBTs, silicon carbide and gallium nitride components, modules, and integrated circuits—posted a gross margin of 22.17 percent in 2025 and 22.91 percent in the first half of 2026.¹² By comparison, the Manufacturing and Services division—comprising contract foundry, packaging, testing, and photomask operations—achieved gross margins of 31.97 percent and 34.16 percent over the same periods.¹²

This margin gap directly contradicts common market assumptions that branded device manufacturing yields higher profitability than contract manufacturing.

CR Micro's branded product lines—despite incorporating advanced silicon carbide and integrated device manufacturing capabilities—generate roughly twelve percentage points lower gross margin than contract manufacturing across the same fabrication infrastructure. This divergence is not an isolated event: in 2024, branded products generated a 21.26 percent gross margin compared to 34.34 percent for manufacturing services.¹²

This structure marks a complete reversal from the pre-listing era, when the proprietary device business achieved above-average margins while foundry operations trailed industry peers.⁶ That shift indicates that domestic market competition intensified more rapidly in branded power devices than in specialized foundry, packaging, and photomask services.

Whatever pricing power the silicon carbide and high-end device portfolio holds, that advantage has not yet translated into higher segment gross margins.

The underlying drivers of this margin structure are evident across the industry supply chain. Rapid domestic capacity additions in silicon carbide substrates and epitaxy during 2023 and 2024 outpaced immediate device demand, driving down input costs and finished device prices. Simultaneously, automotive procurement practices emphasize multi-sourcing, meaning a design win with an electric vehicle manufacturer secures an opportunity to bid rather than long-term pricing power. Furthermore, CR Micro competes within a crowded field of domestic suppliers targeting EV charging, solar inverter, and commercial vehicle markets. The company's own disclosures reflect these sector-wide pricing pressures, highlighting in its 2026 interim report that its photomask division faced intensified price competition from new entrants alongside declining tape-out volumes from upstream fabrication plants.²

A notable contributor to reported earnings in 2026 stemmed from an investment holding rather than device sales. CR Micro held an equity stake in Xiamen-based silicon carbide epitaxial wafer manufacturer Epiworld International (瀚天天成电子科技), which completed an initial public offering on the Hong Kong Stock Exchange on March 30, 2026 at HK$76.26 per share.¹ The resulting investment revaluation generated RMB 267 million in fair-value and financial asset disposal gains within first-half non-recurring items, complemented by RMB 97 million in government subsidies.² Total non-recurring items reached RMB 308 million out of RMB 722 million in reported net profit for the half-year—explaining why headline net profit doubled while net profit excluding non-recurring items grew by 51 percent.²

The calibrated conclusion. Characterizing silicon carbide as an immediate, high-margin earnings driver is not supported by financial disclosures. The evidence demonstrates that CR Micro has established a technically competitive wide-bandgap portfolio and generated rapid revenue growth from a small baseline, but has yet to translate that revenue into expanded gross margins. The primary metric to monitor remains the Products and Solutions segment gross margin in future financial filings. If silicon carbide and advanced silicon products possess structural pricing power, that margin figure must narrow the gap with the Manufacturing and Services division over time. A modest one-percentage-point gain since 2024 indicates initial movement, but not a structural shift.

The second optionality bet: smart sensors and intelligent control. The company's second strategic initiative focuses on smart sensors and integrated control platforms. Strategically, this initiative addresses the vulnerability of discrete power components to price commoditization by combining control logic, gate drivers, and power switches into integrated system solutions.

CR Micro's product portfolio in this segment includes microcontrollers fabricated on eight-inch and twelve-inch wafers, power management integrated circuits, battery management chips, gate drivers, and automotive-grade specialty devices. Technical milestones include multiple automotive chips achieving AEC-Q100 qualification, a full-band in-car radio tuner IC entering ten tier-one supplier networks following AEC-Q100 certification, and an ultrasonic radar driver and signal processor receiving a national innovation award.¹ In manufacturing services, underlying specialty process technologies—such as advanced digital BCD, high-precision CMOS, 1,200-volt high-voltage ICs, and a single-chip CMOS-MEMS inkjet printhead platform—reached volume production. Consequently, industrial and automotive applications expanded to 27.2 percent of wafer manufacturing revenue.¹

This shift toward industrial and automotive applications represents a primary channel for specialty foundries to reduce exposure to price-sensitive consumer electronics markets.

However, qualification milestones and supply-chain inclusions reflect commercial access rather than guaranteed revenue or profit margin. As demonstrated during the 2016 to 2018 period, CR Micro has historically expanded technical capabilities and top-line revenue more rapidly than operating margin.⁶ Accordingly, the smart control portfolio represents a viable technical foundation with long-term potential, rather than an established driver of near-term earnings growth.

↑ Back to Top

VIII. Deep Dive: Competitive Landscape, Moat & Frameworks

CR Micro operates across three distinct competitive fronts, holding a stable market position without asserting clear dominance over any group.

The segment picture. In 2025, main-business revenue split roughly 56 to 44 percent between divisions: Products and Solutions generated RMB 6.03 billion, while Manufacturing and Services contributed RMB 4.79 billion.¹ Geographic revenue remains heavily concentrated at home, with domestic sales reaching RMB 9.68 billion compared to RMB 1.14 billion overseas.¹ Within manufacturing services, packaging and testing revenue expanded 24 percent, advanced panel-level packaging grew 14 percent, and photomask revenue surged 40 percent, supported by high-end line utilization exceeding 90 percent and yields above 98 percent.¹ However, sales-channel data reveals an underlying margin headwind: direct sales yielded a 31.55 percent gross margin in 2025, whereas distributor sales returned 18.21 percent, and distributor revenue grew faster.¹ Because distributor sales expanded at a faster pace, the channel mix shift continues to weigh on overall gross margin.

The global incumbents. Global leader Infineon Technologies illustrates the structural distance between CR Micro and premier international peers. For its fiscal year ended September 30, 2025, Infineon reported €14.66 billion in revenue—down 2 percent—alongside €2.56 billion in segment result, representing a 17.5 percent segment margin, while guiding its fiscal 2026 adjusted gross margin to the low forties in percentage terms.⁷ By comparison, CR Micro posted a gross margin of 26.6 percent in the first half of 2026.²⁷ This roughly fifteen-percentage-point margin gap highlights the difference in product mix and pricing power separating global market leaders from China's largest domestic MOSFET producer, even as Infineon management notes growing domestic Chinese competition in IGBTs and high-voltage components.

Infineon is also pursuing AI data center power infrastructure from a position of scale, raising its fiscal 2026 revenue target for AI power solutions to approximately €1.5 billion.⁷ That target alone exceeds CR Micro's total annual gross profit in recent years, illustrating the resource disparity at the high end of the market.

The domestic peers. Among domestic competitors, Hangzhou-based Silan Microelectronics provides the most direct comparison as China's other major power IDM. In 2025, Silan generated RMB 13.05 billion in revenue—up 16.3 percent—and RMB 399 million in net profit attributable to shareholders, up 81.3 percent.¹⁹ While Silan has surpassed CR Micro in total sales volume, it generates lower absolute net earnings. Silan is also expanding aggressively: in October 2025, it announced a RMB 20 billion twelve-inch specialty analog facility in Xiamen targeting 45,000 wafers per month under an IDM model, with construction starting in early 2026 and initial production scheduled for late 2027.²⁰

This parallel expansion underscores a core structural risk: both leading domestic power IDMs are deploying massive capital into twelve-inch specialty capacity aimed at identical automotive, industrial, robotics, and server applications. This synchronized investment heightens the risk of domestic overcapacity.

The comparison also highlights CR Micro's balance-sheet structure. While Silan has funded its expansion directly through debt and consolidated assets, CR Micro relies on minority joint-venture stakes to limit direct leverage. This approach does not eliminate capital intensity risks; it redistributes financial commitments among the parent enterprise, municipal partners, and lenders.

With Hua Hong Semiconductor competing in specialty contract manufacturing, StarPower Semiconductor leading in IGBT modules, and Yangjie Technology expanding in discrete components, China's mature-node power sector remains exceptionally crowded.

The geopolitical wildcard. Geopolitical friction presents a dual operational dynamic, as demonstrated by Wingtech Technology and its Dutch subsidiary Nexperia. In October 2025, the Netherlands invoked its 1952 Goods Availability Act to assume control of Nexperia, citing governance shortcomings and economic security.²¹ Beijing responded by restricting export shipments from Nexperia's mainland facilities, disrupting European automotive supply chains. Wingtech reported a 2025 net loss of RMB 8.7 billion due largely to the loss of control, subsequently filing an RMB 8 billion damage suit in a Chinese court in May 2026.²¹

For CR Micro, the incident carries two distinct implications. While Nexperia's restructuring accelerated domestic customer qualification of local components, it also highlighted the vulnerability of Chinese semiconductor companies to international policy actions. CR Micro explicitly identified this exposure in its 2026 interim report, noting that expanding multilateral export controls from the United States, Japan, and Europe have reduced supply stability and extended lead times for key manufacturing equipment and high-end consumables.² Because CR Micro relies on international equipment vendors for its twelve-inch wafer fabs, tool procurement remains vulnerable to external policy shifts.

7 Powers, honestly applied. Evaluating CR Micro through Hamilton Helmer’s 7 Powers framework shows that three powers are partially present, though none provides a structural moat.

Scale economies exist operationally but offer limited pricing power. A twelve-inch wafer provides roughly 2.25 times the usable area of an eight-inch substrate, lowering unit processing costs at full yield. However, scale yields a strategic moat only if competitors cannot replicate it. With Hua Hong operating mature twelve-inch lines and Silan constructing equivalent capacity, twelve-inch manufacturing represents industry parity rather than an exclusive cost advantage.

Process power—accumulated manufacturing know-how in trench MOSFETs, super-junction architectures, and specialty BCD processes—represents CR Micro’s most defensible attribute, supported by its photomask, BCD and MEMS platform disclosures.¹ To function as a true power, however, process advantages must yield premium pricing. The compressed gross margins in Products and Solutions indicate that process know-how has not insulated the company from sector-wide price competition.

Counter-positioning is often cited in management narratives, which argue that fabless competitors cannot replicate the iteration speed of an integrated device manufacturer. While valid for complex, high-performance components, this advantage narrows in commoditized segments. In 2025, consumer electronics accounted for 34 percent of Products and Solutions demand, compared to 12 percent for industrial applications and 10 percent communications.¹ By the first half of 2026, the end-market mix shifted toward broad new-energy applications at 37 percent, smart terminals at 22 percent, appliances including smart home at 21 percent, industrial equipment at 12 percent, and emerging fields 8 percent.² Although this shift increases exposure to industrial customers, consumer-oriented demand remains a substantial share of sales.

Switching costs provide partial protection in qualified accounts. Once an automotive or industrial customer qualifies a power component—a qualification cycle often taking up to a year—replacing that device involves qualification expenses. However, this friction protects market share rather than unit pricing. Customer purchasing departments routinely utilize dual-sourcing policies during annual contract negotiations to extract price concessions. Consequently, switching costs slow market share erosion but do not protect gross margins.

Cornered resources, network effects, and brand power do not contribute meaningfully to the business.

The overseas question. Expanding international sales offers a potential avenue to diversify away from domestic price competition. In 2025, CR Micro generated RMB 1.14 billion in overseas revenue at a 25.55 percent gross margin—slightly below its domestic margin of 26.62 percent—indicating that the export business currently functions as a volume outlet rather than a premium channel.¹ Management established a dedicated overseas business development department in 2026, prioritizing exports through its Hong Kong platform in the near term and Southeast Asia over the medium term, using a combination of direct sales and regional distributors.¹³ Representing approximately one-tenth of total revenue, international operations remain an early-stage initiative rather than an immediate financial catalyst.

Porter's Five Forces, honestly applied. A structural evaluation of the industry environment confirms several persistent headwinds.

Supplier power remains elevated and continues to tighten due to international trade restrictions. Equipment procurement and technical support depend heavily on global tool vendors, limiting buyer leverage.

Buyer power is strong. Major appliance manufacturers and automotive suppliers enforce multi-sourcing strategies. Furthermore, faster sales growth within CR Micro’s lower-margin distributor channel highlights intensifying price transparency across end markets.

Rivalry is intense. Domestic IDMs and merchant foundries are adding mature-node and specialty capacity simultaneously into the same domestic market.

Threat of substitution manifests internally as well as externally. The transition from legacy silicon devices to wide-bandgap silicon carbide requires CR Micro to finance new manufacturing platforms while managing potential cannibalization of its established product lines.

Threat of new entry is bifurcated: high capital requirements restrict new wafer fab construction, but low entry barriers for fabless design startups maintain downward pricing pressure in commodity device segments.

The combined framework reveals an enterprise with substantial manufacturing scale and established technical execution operating within challenging industry economics. While cyclical recoveries can offer tactical investment opportunities from valuation troughs, structural capital intensity and market competition have historically constrained multi-cycle returns on equity.

↑ Back to Top

IX. Bull vs. Bear Case & Investor Stress Test

The stress test: the SOE CapEx Trap.

Consider an activist investor evaluating CR Micro's board. The central question does not concern technical capability, but capital allocation: is the enterprise deploying shareholder capital to generate commercial returns, or executing state industrial policy at the expense of equity yields?

The financial record highlights several structural concerns. Return on equity has declined across four consecutive years from its 2021 peak, dropping to 2.91 percent in 2025—a level well below standard cost of equity thresholds.⁹¹ Roughly three-quarters of reported net profit in 2025 was absorbed by equity-method losses from two non-consolidated joint-venture fabs.¹ Financed alongside municipal capital in Chongqing and Shenzhen to support national self-reliance goals, these associate losses expanded during the first half of 2026 despite broader operating recovery.² Non-operating items—specifically government grants and a mark-to-market gain from a supplier's Hong Kong listing—accounted for half of the first-half profit growth.² Meanwhile, research and development intensity contracted from 10.57 percent of revenue in 2025 to 8.57 percent in the first half of 2026, as absolute R&D spending fell 4.15 percent, missing internal full-year targets as acknowledged in risk disclosures.¹² Concurrently, the board committed additional capital to a new industrial fund alongside a related-party asset manager.¹⁸

The counter-argument centers on financial stability. The joint-venture model limits direct balance-sheet liabilities: by mid-2026, consolidated equity reached RMB 23.5 billion against total assets of RMB 31.2 billion, supported by RMB 8.96 billion in cash and equivalents, while short-term borrowings dropped 73 percent as bank debt was retired.² Operating cash flow expanded 95 percent during the half to RMB 1.39 billion.² Rather than a distressed enterprise facing debt pressures, the company operates as an under-earning business backed by a strong balance sheet. Cash dividends continued throughout the downturn, the controlling shareholder avoided diluting minority equity, and the twelve-inch manufacturing capacity aligns with long-term domestic substitution goals across automotive and artificial intelligence power systems.

The structural assessment remains inconclusive: the capital trap risk is neither fully established nor entirely disproved, leaving the burden of execution on management. Four consecutive years of declining returns require concrete operational results rather than forward projections.

Beyond capital allocation, governance disclosures present a second analytical challenge. CR Micro releases monthly investor relations records detailing operational metrics—such as wafer capacity utilization, order visibility windows, and price adjustment timelines—yet provides limited guidance on financial return targets.¹³ The company publishes no explicit group return-on-capital benchmarks, minimum hurdle rates for twelve-inch capital deployments, or targeted breakeven dates for joint-venture facilities. For an enterprise defined by capital deployment, this lack of return guidance remains notable.

Executive incentives present a third factor. Senior management compensation is governed by central state-owned enterprise guidelines, with the board adopting updated remuneration rules on a trial basis in May 2026.¹⁸ While state compensation caps mitigate speculative risk-taking, they also loosen the direct link between executive remuneration and per-share financial returns. With China Resources holding a 66.21 percent controlling stake alongside national policy priorities, corporate incentives prioritize operational alignment over short-term minority return optimization.

These dynamics highlight an enterprise where strategic direction and minority shareholder returns align in general orientation, but diverge in execution priority—a key consideration at the bottom of a capital-intensive industry cycle.

The bull case.

First, import substitution provides a substantial addressable market. China represented $29.1 billion of the $75.5 billion global power semiconductor market in 2025—or 38.6 percent—with market research firm Omdia projecting domestic demand to expand to $38.1 billion by 2028.¹ Domestic suppliers control a minority share of high-value segments; server power management localization, for example, remained between 10 and 20 percent in 2025.¹ Overseas supply disruptions, including the Nexperia sanctions, have further encouraged domestic manufacturers to qualify local suppliers.

Second, operational demand is recovering alongside price adjustments. CR Micro issued price increase notifications in February and July 2026, informing investors in late July of potential additional adjustments if fab capacity remained constrained and raw material costs escalated.¹³²² Management reported that customer order visibility stretched up to nine months, pushing delivery backlogs to record levels.¹³ Interim financial disclosures confirmed near-capacity fab utilization rates.²

Third, fixed-cost operating leverage enhances margin expansion during cyclical upturns. In the first half of 2026, revenue increased 17.4 percent while blended gross margin rose roughly one percentage point to 26.6 percent, driving operating profit expansion above top-line growth.² Segment sales in Products and Solutions expanded 23 percent, wafer manufacturing rose 10 percent, and packaging and testing grew 16 percent.² If joint-venture facilities reduce operating losses while core fabs maintain high utilization, profitability could expand rapidly.

Fourth, state backing offers financing advantages. Central state-owned enterprise status under SASAC ensures access to debt capital, municipal co-investment, and policy support unavailable to non-state competitors.

The bear case.

First, industry-wide capacity expansion presents risk of structural overcapacity. Multiple domestic power semiconductor producers are simultaneously expanding mature-node and specialty analog manufacturing lines targeted at identical domestic end markets, reproducing the supply dynamics that led to sector price deflation between 2022 and 2025.

Second, depreciation costs remain substantial and fixed across multi-year horizon ramps. Equity-method associate losses represent only a portion of this burden, as consolidated depreciation across packaging, photomask, and fab infrastructure directly increases cost of goods sold. Subsidiary disclosures reflect this pressure: the Chongqing advanced packaging entity, 华润润安 Run'an, with RMB 2.42 billion in registered capital, recorded a net loss of RMB 93 million in the first half of 2026, while photomask subsidiary 无锡迪思 Wuxi Disi shifted from an RMB 22 million net profit in 2025 to an RMB 18 million net loss in the first half of 2026 as high-end equipment depreciation coincided with price competition.¹²

Third, multilateral trade restrictions constrain equipment procurement and servicing. As acknowledged in corporate risk disclosures, export controls limit access to manufacturing tools, replacement parts, and technical support.² Rather than leading-edge lithography, equipment risks center on mid-tier processing tools—including deposition, etch, ion implantation, and metrology systems—alongside specialized consumables. Inability to secure replacement parts directly threatens fab output, with reported tool delivery delays providing an early indicator of supply chain friction.

Fourth, policy-driven customer demand introduces concentration risk. A substantial portion of sales volume relies on domestic customers transitioning from foreign components under national industrial directives. While providing strong near-term volume, policy-driven purchasing can shift rapidly toward competing domestic vendors or alternative technologies, offering less commercial defense than proprietary product moats.

Fifth, earnings quality depends heavily on non-operating items. Non-recurring income—comprising government subsidies, fair-value adjustments, and asset disposal gains—represented over 40 percent of reported net profit in the first half of 2026.² Investment revaluations carry downside exposure: the Epiworld equity holding that boosted first-half earnings remains subject to a one-year lock-up following its March 2026 listing, meaning subsequent share price fluctuations will directly impact reported net income.¹

The three key performance indicators.

Evaluating CR Micro's long-term financial trajectory relies on three primary metrics disclosed in regular reporting.

First: net earnings performance of the two twelve-inch joint ventures. Disclosed separately in annual and interim reports, equity-method results from the Chongqing and Shenzhen facilities measure whether the company's major capital deployment generates positive economic returns. Narrowing losses alongside revenue growth indicate progress toward breakeven, whereas expanding losses during industry recoveries highlight persistent fixed-cost drag.

Second: Products and Solutions segment gross margin relative to Manufacturing and Services. This metric tests the competitive moat of the integrated device model and silicon carbide product portfolio. If proprietary device integration delivers pricing power, branded product margins should narrow the eleven-percentage-point gap with manufacturing services and eventually exceed foundry margins.

Third: net profit excluding non-recurring items and corresponding return on equity. Because headline net income includes government subsidies and investment revaluations, net profit excluding non-recurring items (扣除非经常性损益后的净利润) provides the clearest measure of core operating performance and capital returns relative to cost of capital.

↑ Back to Top

X. Epilogue & Playbook Lessons

A single hypothetical photograph captures the corporate arc: a 1997 Wuxi cleanroom producing six-inch wafers on obsolete equipment alongside a 2026 Chongqing cleanroom running twelve-inch wafers on modern tools—funded by municipal state capital, owned 19 percent by a public company whose equity holders bear the initial losses, and selling wafers directly back to the listed parent.

Nearly three decades apart, the fundamental strategic question remains unchanged: can Chinese state capital, deployed at scale into fixed-cost manufacturing assets, generate economic returns before technology cycles and market pricing erode profitability?

Four primary lessons emerge from this corporate record.

The IDM argument in power semiconductors is valid but bounded. Process and design co-optimization builds genuine technical capability where device physics and manufacturing processes intersect. However, an integrated device model does not guarantee gross margin protection during industry downcycles—a dynamic demonstrated by the 2022-to-2025 profit decline—nor has it translated into premium pricing for branded devices relative to contract foundry services. The integrated approach represents a technical capability that has yet to convert into a durable economic moat.

Listing venue creates valuation expansion, but venue choice should not be confused with operational skill. CR Micro's 2004 Hong Kong initial public offering raised HK$342 million and left the enterprise with billions of renminbi in accumulated balance-sheet losses, whereas its 2020 STAR Market listing raised more than twelve times as much and expanded its valuation by an order of magnitude in a single trading session. These outcomes reflect capital market conditions and domestic policy support rather than fundamental changes in business quality. Interpreting policy-driven valuation expansion as management brilliance misreads the evidence, especially as the stock's wide twelve-month trading band demonstrates how quickly market multiples can adjust.

Operating leverage operates symmetrically, and minority equity stakes do not eliminate fixed-cost drag. Structuring the Chongqing and Shenzhen twelve-inch fabs as non-consolidated joint ventures keeps construction debt off CR Micro's balance sheet, but does not insulate its net profit. Whether fixed depreciation costs flow through cost of goods sold or appear below the line as a share of joint-venture losses, equity investors absorb the financial burden. Ultimate profitability depends on sustaining high wafer throughput and commercial yields.

Market information asymmetry stems from translation and focus rather than a lack of public disclosure. The empirical evidence needed to evaluate CR Micro's business model is publicly available in Chinese regulatory filings: associate profit-and-loss tables, segment gross margin breakdowns, direct versus distributor channel dynamics, non-recurring subsidy schedules, and monthly investor relations transcripts. Investors who examine these primary disclosures can evaluate corporate progress without relying on high-level promotional narratives.

CR Micro in 2026 stands as an established industrial enterprise with a strong domestic market footprint, a conservative balance sheet, credible process engineering, and improving cyclical demand supported by product price adjustments. Yet it concurrently carries a four-year history of contracting returns on equity, substantial unproven twelve-inch capital commitments, branded-product gross margins that lag its own foundry division, and recent senior leadership transitions during a pivotal operational expansion. It is neither the invincible national champion portrayed in bullish market commentary nor the inefficient state vehicle depicted by critics. Instead, it is a capital-intensive manufacturer executing a high-stakes expansion, offering detailed disclosures that allow attentive market participants to track its execution in real time.

From a loss-making 1990s state fab operating five generations behind global leaders, CR Micro has traveled a notable operational distance. Whether its ongoing twelve-inch capital deployment delivers acceptable returns for public equity holders remains the central question that upcoming financial disclosures from its associate ventures will address.