Nexchip Semiconductor: The Display Foundry Champion's Crucible

I. Introduction & The Hefei Semiconductor Bet

On a July morning in 2026, a Chinese chipmaker little known to most Western investors listed on the Hong Kong Stock Exchange, raising roughly US$890 million. Heavy oversubscription pushed pricing to the top of the range at HK$32.30 per share for 216.17 million H shares, with twenty cornerstone investors—ranging from Hillhouse to a state-owned automaker—locking up nearly half the deal before public trading opened.¹² The company was 合肥晶合集成 Nexchip Semiconductor. Eleven years earlier, its site had been farmland on the eastern edge of Hefei, an inland Anhui capital better known for tofu and freight rail than silicon.

How that farmland became the world's largest foundry for display driver chips offers one of the most instructive case studies in modern Chinese industrial policy—and a revealing test of what "market leadership" is actually worth in commodity manufacturing.

The central paradox can be stated simply: Nexchip is simultaneously the undisputed global leader in its core product category and a company whose 2025 net profit margin hovered around four percent.³

Examining both sides of that dynamic clarifies the picture. In 2025, Nexchip generated ¥10.885 billion in revenue, up 17.69% year on year, while net profit attributable to shareholders rose 32.16% to ¥704 million, producing a gross margin of 25.52%.³⁴ By the first quarter of 2026, it climbed to eighth place in TrendForce's global pure-play foundry ranking—its highest position ever—and third among mainland Chinese foundries.⁵ By both internal and third-party metrics, it is the world's largest foundry for 显示驱动芯片 display driver ICs (DDIC), the unglamorous components that sit at the edge of LCD and OLED panels to translate digital video signals into precise pixel-glowing voltages.⁶

Yet that eighth-place global ranking represented roughly US$400 million in quarterly revenue and just a 0.8% share of the global foundry market, compared to 台积电 TSMC's 72% share in the same quarter.⁵ The entire mature-node cohort—Nexchip, 世界先进 Vanguard International Semiconductor, 力积电 PSMC, and Tower—clustered around eight-tenths of one percent each, separated only by rounding errors. Leading the display driver sector is akin to being the world's largest manufacturer of a single ball-bearing size: technically demanding and operationally impressive, but fundamentally distinct from pricing power.

That distinction forms the core of this narrative. Nexchip serves as a prime test case for China's semiconductor localization drive—the strategy of replacing imported volume chips with domestic production. The company executed the industrial-policy playbook with notable speed: constructing fabs faster than almost any competitor, capturing market share from Taiwanese incumbents, and achieving scale within a decade. What remains unproven is whether this operational expansion translates into durable value for minority shareholders.

The financial rationale for skepticism appears on the company's income statement. In 2025, Nexchip recorded depreciation and amortization of roughly ¥4.1 billion against ¥10.9 billion in revenue, meaning depreciation alone absorbed nearly forty cents of every revenue yuan while growing faster than top-line sales.[^7] Financially, a twelve-inch wafer fab operates as a massive fixed-cost structure with a minor variable-cost component. When wafer prices remain firm and capacity is fully utilized, operating leverage boosts earnings significantly. When prices weaken, that same leverage works against profitability with equal speed.

Recent financial results illustrate this dynamic directly. In the first half of 2026, Nexchip increased revenue by 12.40% to ¥5.766 billion, even as net profit attributable to shareholders dropped 26.12% to ¥245 million.⁷ Revenue growth coupled with falling profits is the classic hallmark of a business lacking pricing authority.

Understanding the economics of a "pure-play foundry" requires examining how the model functions in practice. Nexchip designs no products, owns no consumer brands, and holds no proprietary claim to the end chips produced in its facilities. Instead, it owns advanced manufacturing facilities and sells processing capacity, measured in silicon wafers containing hundreds of patterned chips. Fabless customers supply the chip designs, Nexchip manufactures finished wafers, and customers slice them into individual chips for sale. Consequently, foundry revenue equals wafer shipment volume multiplied by wafer price, while costs are dominated by equipment depreciation alongside electricity, chemicals, and labor.

This business structure carries a major strategic consequence. Because customers retain design ownership while the foundry owns only the process, customers can shift production to competing foundries. A pure-play foundry derives leverage not from proprietary product lock-in, but from offering superior cost, speed, or yield at a given point in time. This dynamic explains why extraordinary returns in the foundry industry concentrate in a single leader operating at the advanced technological frontier, where substitute capacity does not exist. Nexchip operates outside that frontier, in market segments where alternative capacity is plentiful.

Four central themes structure this analysis. The first is the mechanics of the 合肥模式 "Hefei model"—municipal venture capitalism deployed on an industrial scale—and how state ownership shapes corporate incentives. The second is the single-crop risk: the vulnerability created by relying heavily on a single product category whose customers can readily shift order volumes. The third is the historical erosion of mature-node moats, showing how volume market share alone fails to protect operating margins. The fourth is the node transition: the capital-intensive attempt to move from 110-nanometer LCD drivers into 55, 40, and 28-nanometer process nodes, as well as image sensors, power management, and automotive silicon—a shift funded by cash flows from the legacy business it aims to diversify beyond.

Understanding this trajectory requires examining how a municipal government set out to establish a domestic semiconductor industry from scratch.

↑ Back to Top

II. Cross-Strait Origins & The "Hefei Model" (2015–2017)

In 2015, China assembled more display panels than any other country in the world. Yet it imported the overwhelming majority of the driver chips required to operate those panels. While the display panels were made domestically, the silicon at their edges was not. For a municipal government that had staked its economic strategy on the display sector, this represented a critical supply chain gap—and an opportunity.

Hefei had a track record with this playbook. The city's state investment arm, 合肥建投 Hefei Construction Investment Holding Group, had accomplished a feat rare among Chinese local-government finance vehicles: executing a transformative industrial strategy successfully. Hefei had backed 京东方 BOE Technology's display fabs when BOE was a struggling underdog, helping it become one of the world's two largest panel manufacturers. The city later applied the same approach to electric vehicles with 蔚来 NIO. This model—using the municipal balance sheet as an anchor investor to build an industrial cluster—became known in China as the 合肥模式 Hefei model, turning a provincial capital into a key allocator of industrial capital in Asia.

The missing piece in Hefei's display cluster was the driver chip. In April 2015, the municipal government signed a project agreement with a Taiwanese partner. In May of that year, 合肥晶合集成电路 Nexchip was formally established as a joint venture between Hefei Construction Investment and 力晶创投 Powerchip Venture Capital, the investment holding vehicle of Taiwan's 力晶科技 Powerchip.⁸⁹

Powerchip was a strategic choice of partner. Founded in 1994, Powerchip had spent its first 15 years as a DRAM manufacturer before being hit severely by the memory downturn that followed the 2008 financial crisis. It subsequently pivoted into a pure-play foundry model, offering its manufacturing capacity to fabless chip designers rather than risking its balance sheet on commodity memory pricing.⁹ That pivot provided Powerchip with mature-node, high-voltage process technology on twelve-inch wafers, though it lacked the capital to fund greenfield expansion. Hefei presented the inverse profile: abundant capital and land, but no proprietary semiconductor technology.

The resulting deal was a technology-for-capital swap. In 2015, Powerchip provided Nexchip with foundational process specifications for 150-nanometer, 110-nanometer, and 90-nanometer display driver manufacturing. In October of that year, Powerchip and Hefei Construction Investment signed a formal agreement outlining the execution details for the three process transfers.⁸ Construction of the first-phase fab—a ¥12.81 billion project in Hefei's 新站高新区 Xinzhan High-Tech Zone—began the same month.⁹ The initial ownership structure gave Hefei Construction Investment 31.14%, Powerchip 27.44%, and the 合肥芯屏 Hefei Xinping industrial investment fund 21.85%.¹⁰

The plan incorporated one distinctly contrarian operational choice. At the time, nearly all display driver chips globally were produced on eight-inch wafers using older, fully depreciated manufacturing lines kept operational because the production costs were low. Nexchip instead built a twelve-inch line dedicated to display drivers. A twelve-inch wafer offers roughly 2.25 times the surface area of an eight-inch wafer, yielding significantly more chips per wafer and superior unit economics—provided the fab operates near full capacity. It was a calculated bet that display driver demand was sufficiently large and growing fast enough to absorb an unprecedented scale of output in that chip category.

Overseeing the operational build was 蔡国智 Tsai Kuo-Chih, a Taiwanese semiconductor executive born in 1953 who served as and remains the company's chairman.¹¹ The resulting governance structure proved effective during the ramp-up phase: Taiwanese manufacturing discipline managed fab operations, Hefei state capital absorbed early financial losses, and a municipal government with a long investment horizon held the equity.

Those financial losses were substantial. Nexchip's first-phase fab began production in October 2017 and achieved a capacity of 10,000 wafers per month by December 2018.⁹ In 2018, the company generated ¥218 million in revenue while posting a net loss of ¥1.19 billion. The following year, revenue rose to ¥534 million, but net losses widened to ¥1.24 billion.[^7] The fab was burning more than two yuan in losses for every yuan of revenue it brought in. While standard commercial investors with typical cost-of-capital constraints would likely have declined to fund such operating losses, a municipal government focused on building an industrial ecosystem was willing to absorb them.

Testing the "Hefei ecosystem moat"

A primary argument for Nexchip's early strategic positioning is that proximity to BOE created a cornered customer relationship characterized by guaranteed order volumes and a captive local supply chain. This hypothesis warrants rigorous examination, as it underpins much of the ecosystem-moat narrative referenced by market observers.

Structural realities in the display driver integrated circuit (DDIC) value chain challenge this assumption. BOE does not purchase driver chips directly from Nexchip. Instead, BOE buys finished driver chips from fabless design houses—including 联咏科技 Novatek, 集创北方 Chipone, 奕力科技 Ilitek, and 奇景光电 Himax—which independently determine where to allocate their manufacturing contracts.¹² These design houses historically distribute production across foundries such as 台积电 TSMC, 联电 UMC, 世界先进 VIS, 中芯国际 SMIC, and Nexchip, reallocating volumes based on pricing, capacity availability, and lead times. Physical proximity to panel manufacturing facilities shortens supply logistics and accelerates feedback loops on yield or technical specifications, but it does not guarantee customer volume or grant pricing power.

The Hefei relationship is therefore better understood as providing demand visibility and regional infrastructure advantages rather than a locked-in customer base. This represents a real but bounded operational benefit. The distinction is critical: a foundry with a captive customer base is insulated from pricing cycles, whereas a foundry with a convenient location remains exposed to broader market price competition.

By the end of 2019, Nexchip possessed an operational fab, process technology licensed from Taiwan, a patient municipal shareholder, and billions of yuan in accumulated losses. What the company required next was volume demand, which global market conditions provided in 2020.

↑ Back to Top

III. The Display Driver Supercycle & Scale Ramp (2018–2022)

In the spring of 2020, purchasing managers across the electronics supply chain suddenly found display driver wafers impossible to buy at any price. Every laptop, monitor, television, and tablet demanded by a locked-down global population required driver chips. Yet those chips were manufactured on mature, low-margin, capacity-constrained lines that foundries had neglected to expand for years because unit economics never justified the capital expenditure. The primary bottleneck for the entire consumer electronics industry turned out to be one of its most unglamorous components.

Nexchip had spent the previous five years building precisely the capacity the market now lacked.

The subsequent operational ramp was remarkably fast. Monthly capacity reached 25,000 wafers in July 2020, crossed 40,000 in March 2021, and expanded beyond 100,000 wafers per month by March 2022—a fourfold scale-up in under two years in an industry where planning and constructing a single fab typically requires three years.¹³ Concurrently, the foundry advanced its internal manufacturing capability beyond its licensed 110-nanometer baseline down to 90 nanometers, having built a proprietary touch-and-display driver integration platform in 2018 on top of Powerchip's 90-nanometer process technology.⁹

Financial performance mirrored this expansion curve. In 2020, Nexchip generated ¥1.51 billion in revenue but posted a gross loss of ¥146 million, as wafer production costs still exceeded selling prices. By 2021, revenue surged to ¥5.43 billion as gross profit turned positive at ¥2.45 billion. In 2022, top-line sales reached ¥10.05 billion and net profit reached ¥3.05 billion.[^7] Over that two-year period, gross margin expanded from negative territory to approximately 43%, supported by an 85.9% capacity utilization rate in 2022.¹²

This earnings trajectory was driven primarily by external market conditions rather than fundamental technological differentiation. Nexchip operated substantially the same fab infrastructure and process technology in 2022 as it had in 2020, selling to a largely identical customer base. The dramatic margin expansion resulted from a global supply deficit that lifted both wafer prices and shipment volumes. While operational execution allowed Nexchip to capitalize on the surge, industry-wide tightness—rather than proprietary pricing power—created the windfall.

By 2022, Nexchip had converted its production expansion into top-tier market share. The company established itself as the world's largest pure-play foundry for display driver ICs and climbed into the global top ten among pure-play foundries—becoming the first mainland Chinese firm built from scratch within seven years to displace entrenched Taiwanese incumbents at the top of the display driver category.⁶⁹ Display drivers accounted for 71.2% of its revenue that year.¹²

Testing the "structural high-margin moat"

The 2022 financial peak prompted claims that Nexchip had secured a structural, high-margin competitive moat based on scale economies in mask sets, cumulative yield learning, and customer switching costs. However, company disclosures present three clear counter-arguments to this narrative.

First is customer concentration. Nexchip's top five customers consistently generated over 60% of revenue—accounting for 60.7%, 64.2%, 62.2%, and 58.1% across the reporting periods in its Hong Kong listing documents—with Taiwanese design house 联咏科技 Novatek alone contributing roughly 61% of total revenue at its peak before declining to 51% and eventually 31% as customer mix diversified.¹² A foundry relying on a single customer for more than half its sales functions as a price taker rather than a price setter.

Second is product standardization. A 90-nanometer or 110-nanometer display driver is a standardized component produced on high-voltage manufacturing processes available at multiple foundries. Qualifying a secondary manufacturing source requires time and tooling expense for fabless design houses, but it remains a standard operational procedure rather than an insurmountable barrier to entry. This dynamic reflects moderate operational lock-in rather than high switching costs.

Third is the margin collapse following supply normalization. As global foundry capacity rebalanced in late 2022, Nexchip's utilization rate declined from 85.9% to 72.5% within twelve months, while its gross margin dropped from roughly 43% to approximately 20%.¹² A genuine competitive moat shields profit margins during cyclical downturns; a temporary capacity shortage does not.

Historical evidence supports a more bounded conclusion regarding Nexchip's market position. The company maintains a durable volume position in display driver manufacturing, retaining the global top ranking in DDIC foundries for several consecutive years and holding roughly one-quarter of the global DDIC foundry market in 2024.⁶ However, this volume leadership does not translate into structural pricing authority. Nexchip operates a cost-competitive production engine that captures substantial volume during demand peaks, but its realization prices and gross margins remain tightly bound to the broader semiconductor cycle.

Nexchip reached the peak of the 2022 supercycle running at full capacity, holding market leadership in display drivers, and backed by state investors eager to fund further expansion. In 2023, the company turned to public equity markets to finance its next phase of growth.

↑ Back to Top

IV. The Mega STAR Market IPO & Capital Allocation (2023)

Timing an initial public offering involves both strategic execution and market cycle alignment, and Nexchip's 2023 listing highlighted the tensions between the two.

The company listed on the 上海证券交易所科创板 Shanghai Stock Exchange STAR Market on May 5, 2023, under ticker 688249, issuing shares at ¥19.86 each to raise ¥9.96 billion.¹⁴ At the time of pricing, it marked the largest A-share initial public offering of the year to date and the largest listing in the history of Anhui province—representing municipal industrial policy executed on a national venue.¹⁴ By the company's own description, it was also the first pure-play wafer foundry to list on China's domestic equity market.¹³

The listing prospectus outlined an expansionary roadmap: deploy new capital to scale manufacturing capacity and advance to smaller process nodes. Proceeds were targeted at the second and third phases of the Hefei fab complex, expanding planned twelve-inch capacity substantially beyond the 100,000 wafers per month reached in 2022. The capital also backed an R&D program aimed at 55-nanometer and 40-nanometer OLED display drivers, 55-nanometer high-resolution image sensors, and power management process platforms.⁹ For an enterprise that spent its initial eight years focused primarily on display drivers across two process nodes, the IPO signaled a formal effort to diversify beyond its legacy single-product concentration.

Timing, however, proved challenging. Nexchip listed near the peak of the global display driver cycle. By the closing trading days of 2023, the stock fell to ¥17.25—dropping below its ¥19.86 offer price.¹⁴ Public investors experienced immediate capital losses not from operational failure, but because newly funded capacity expanded directly into an industry-wide pricing downturn.

The ownership architecture

The resulting ownership structure defines the company's governance and strategic direction. Hefei's municipal state-asset supervisory body functions as the ultimate controlling shareholder: 合肥建投 Hefei Construction Investment Holding Group holds 23.35% directly and an additional 16.39% indirectly through the 合肥芯屏 Hefei Xinping industrial investment fund, controlling 39.73% of voting rights.¹⁵ Founding partner Powerchip retained a substantial though gradually declining equity stake, while public shareholders hold the remaining float.

This governance structure means capital allocation decisions do not rest solely with a management team accountable to dispersed public shareholders. Instead, strategic choices reflect the priorities of a municipal controlling shareholder whose mandate encompasses local industrial clustering, job creation, and supply-chain security. While local economic development goals often align with commercial success, they can diverge during industry downturns when maintaining high utilization and building regional capacity take precedence over maximizing immediate net profit margins.

Testing the "scale economies through capex" claim

The theoretical rationale for aggressive capital expenditure in wafer foundries is straightforward: twelve-inch semiconductor fabs carry high fixed overheads, meaning higher wafer volumes dilute unit costs and strengthen competitive positioning. Rapid, large-scale construction theoretically compounds this cost advantage.

Empirical cash flow data complicates this thesis, largely due to the timing of capital deployment.

Nexchip spent approximately ¥7.4 billion on property, plant, and equipment in 2023, ¥13.2 billion in 2024, and ¥9.7 billion in 2025.[^7] Over the same period, operating cash flows totaled negative ¥161 million in 2023, ¥2.76 billion in 2024, and ¥3.84 billion in 2025.[^7] Consequently, across three consecutive years following its public listing, cumulative free cash flow remained deeply negative by more than ¥23 billion. Capacity growth was financed primarily through public equity proceeds, government co-investment, and corporate debt rather than self-sustaining operational cash flow.

This heavy capital deployment generates a persistent financial drag through elevated depreciation charges. Annual depreciation and amortization rose from roughly ¥2.8 billion in 2022 to ¥3.24 billion in 2023, ¥3.42 billion in 2024, and ¥4.13 billion in 2025.[^7] Fabs commissioned near the market peak began depreciating directly into a cyclical trough, creating a large, inflexible cost burden that depresses reported earnings regardless of prevailing wafer selling prices.

To manage this capital burden, Nexchip adjusted its financing structure for subsequent expansions. In September 2024, its third-phase project entity, 皖芯集成 Wanxin Integrated, raised ¥9.55 billion in private equity from external investors.⁹ While offloading capital requirements from the listed company's balance sheet, this co-investment structure introduced minority equity holders whose financial return expectations create additional governance considerations.

Evaluating Nexchip's capital allocation reveals a distinct contrast between construction efficiency and market cycle timing. The company demonstrated strong project execution by building manufacturing facilities rapidly and cost-effectively, but commissioned that capacity at a cyclical peak in mature-node pricing. Validating this aggressive expansion strategy requires demonstrating that the added capacity can sustain high utilization and maintain gross margins above the mid-twenty percent range across a full multi-year semiconductor cycle—a benchmark the company has yet to achieve.

Instead, shortly after the listing, mature-node foundries entered one of the sharpest pricing downturns of the past decade.

↑ Back to Top

V. The Great Foundry Downcycle & Margin Meltdown (2023–2024)

The bullwhip effect is a foundational concept in supply chain economics: when end demand softens even slightly, each layer of the supply chain—from retailers and consumer brands to module makers, chip designers, and foundries—cuts orders more aggressively than the tier above it as companies attempt to shed excess inventory accumulated during shortages. By the time that demand signal reaches foundries at the base of the chain, a minor slowdown in consumer purchasing amplifies into a steep downturn.

That dynamic hit Nexchip directly in 2023. Display panel manufacturers curtailed procurement, while fabless driver-chip design houses that had stockpiled wafers discovered they were holding months of unsold inventory. With foundries across mainland China and Taiwan simultaneously bringing online new mature-node capacity constructed during the preceding shortage, intense price competition followed.

The resulting financial results demonstrated how quickly high fixed overhead turns operating leverage against a foundry during a slump. Revenue dropped 27.9% to ¥7.24 billion, gross margin shrank from roughly 43% to around 20%, and capacity utilization slipped to 72.5%. Net profit attributable to shareholders fell more than 90% from ¥3.05 billion the previous year to ¥212 million.[^7]¹²

More revealingly, operating cash flow in 2023 turned negative at minus ¥161 million.[^7] For a capital-intensive manufacturer recording ¥3.2 billion in annual depreciation—a non-cash charge that typically provides a cash-flow buffer beneath net income—failing to generate positive operational cash flow underscored the severity of the contraction. A negative working-capital swing of roughly ¥4 billion drove this outcome,[^7] demonstrating how downcycle pressures on accounts receivable, inventory, and payables can compound the primary impact of falling average selling prices.

Corporate messaging shifted alongside the financials. Through 2022, executive communications highlighted display driver IC leadership, volume dominance, and scale economies. From 2023 onward, corporate strategy focused heavily on product diversification into image sensors, power management ICs, microcontrollers, and logic chips. While expanding the product lineup represents a logical operational adjustment to market conditions, it also highlights the limitations of relying on a single, highly cyclical product category where manufacturers lack pricing power.

A note on how this company communicates

A structural aspect of corporate disclosure shapes how external investors analyze the business. Nexchip does not host open quarterly earnings calls featuring real-time analyst question-and-answer sessions common among U.S.-listed semiconductor firms. Instead, communications rely on statutory filings, regulatory announcements, and the mainland market practice of the 业绩说明会 investor results briefing, where questions are submitted and answered primarily in writing. The company's Hong Kong listing has begun to expand these disclosures by introducing a dual-listing reporting framework targeting international institutional investors.

Consequently, critical details regarding operational health often emerge through financial filings rather than corporate presentations—including updates on IPO proceeds that signal project timelines, disclosures on related-party transactions detailing subsidiary financing terms, and the variance between consolidated net profit and net profit attributable to the parent company. While headline financial results reflect broader industry cycles, evaluating firm-specific performance requires analyzing these underlying mechanics.

The 2024 rebound, and what it did and did not prove

When market demand rebounded in 2024, recovery was swift. In October 2024, Nexchip pre-announced that net profit attributable to parent shareholders for the first nine months would reach between ¥270 million and ¥300 million—a year-on-year increase of 744% to 838%—and noted that manufacturing facilities had operated at full capacity continuously since March of that year.¹⁶ Full-year 2024 revenue reached ¥9.25 billion, up 27.69%, while net profit attributable to shareholders increased 151.78% and gross margin recovered to approximately 25%, supported by an average capacity utilization rate of 94%.¹²

These results warrant careful examination. The 2024 earnings rebound was driven primarily by higher production volumes and asset utilization rather than expanding pricing power. Capacity utilization rose from 72.5% in 2023 to 94% in 2024. In a business with high fixed manufacturing costs, a 21.5-percentage-point increase in utilization automatically expands margins by spreading fixed depreciation costs over more output, regardless of changes in average selling prices. Yet even at 94% utilization, gross margin recovered to only about 25%—roughly 18 percentage points below the 43% peak achieved in 2022. This persistent margin gap reflects ongoing pricing pressure across mature nodes.

A comparison between 2022 and 2024 highlights this structural shift: in 2022, Nexchip achieved a 43% gross margin with 85.9% utilization, whereas in 2024, a higher utilization rate of 94% yielded a gross margin of just 25%. Despite high factory loading, profitability remained constrained by lower unit pricing. This performance indicates that scale, regional integration, and state support did not fully shield the company from industry-wide price erosion. Rather than guaranteeing high profit margins across the cycle, competitive manufacturing costs enabled Nexchip to maintain positive operating income during a severe downturn that forced higher-cost foundries into operating losses.

An additional accounting factor affects reported profitability. Nexchip's net profit attributable to parent shareholders has regularly exceeded total consolidated net profit, indicating that non-controlling minority interests in subsidiary entities absorbed a portion of operating losses. In 2023, net profit attributable to the parent company of ¥212 million exceeded consolidated net profit of ¥119 million.[^7] By 2025, that divergence widened: consolidated net profit totaled ¥467 million, down 3.19%, while attributable net profit reached ¥704 million, up 32.16%.¹⁷³ Consequently, headline net profit growth figures partly reflect how subsidiary operating losses are allocated across corporate entity structures.

By early 2025, Nexchip had navigated its first complete downcycle as a public company, restored factory utilization, and expanded its product portfolio into non-display categories.

↑ Back to Top

VI. Current Management & Strategic Governance

On February 13, 2025, Nexchip announced that general manager 蔡辉嘉 Tsai Hui-chia—one of five individuals formally designated as "core technical personnel"—had resigned for personal and family reasons and would hold no further position at the company.¹⁸ Joined in April 2016 at age 51, Tsai had risen from vice-general manager of operations to executive vice-general manager before taking the top operating role. His 2024 compensation of ¥3.62 million was the highest of any director or executive.¹⁹ His departure reduced the company's designated core technical personnel from five to four.¹⁹

In response, the board asked Chairman Tsai Kuo-Chih, then 72, to assume the general manager's duties on an acting basis.¹⁸ The company stated that the departure would not materially or adversely affect its technological development, core competitiveness, or ongoing operations.¹⁹

While that is standard boilerplate in Chinese market filings, investors must weigh two underlying realities. First, a foundry's competitive edge rests almost entirely on process engineering and operational execution—the tacit knowledge needed to maintain chip yield on high-voltage nodes while ramping new platforms. Second, Nexchip is attempting the most technically demanding transition in its history: moving from licensed process nodes to in-house technology development. Replacing the operating executive who led fab operations during the initial ramp with a 72-year-old chairman serving on an acting basis represents a significant key-person transition.

Below the chairman, executive leadership includes senior vice-general managers 郑志成 Zheng Zhicheng and 邱显寰 Qiu Xianhuan, alongside 朱才伟 Zhu Caiwei, who serves as employee-representative director, board secretary, chief financial officer, and vice-general manager.²⁰

Management credibility: promises against outcomes

Evaluating management requires comparing past operational targets against delivered results.

In October 2024, alongside its earnings pre-announcement, Nexchip stated that its 28-nanometer OLED driver chip would enter volume production in the first half of 2025.¹⁶ By the time the 2025 annual report was published in March 2026, the timeline had shifted: while 40-nanometer high-voltage OLED drivers and 55-nanometer image sensors and logic had reached mass production, and the 28-nanometer logic platform had completed development, the 28-nanometer OLED products remained "under validation."⁴ Because product validation precedes mass manufacturing, the commercial timeline slipped by at least a year.

This delay reflects a broader execution pattern, which is examined further in subsequent sections. It also underscores a persistent valuation question: during its STAR Market listing, the investment thesis relied heavily on peak gross margins achieved during global supply shortages, only for those margins to halve within two reporting periods. While industry price competition was driven by external market conditions, the contrast between IPO-era margin expectations and post-listing performance shapes how public markets evaluate management's forward-looking guidance.

The founding partner heads for the exit

A major governance shift has unfolded with relatively little attention: founding technology partner Powerchip has steadily reduced its equity stake.

In July 2025, Chinese electronics original design manufacturer 华勤技术 Huaqin Technology agreed to acquire roughly 120 million shares from Powerchip Venture Capital at ¥19.88 per share. The ¥2.393 billion cash transaction represented a 6.00% stake, reducing Powerchip's holding from 19.08% to 13.08% and making Huaqin the fourth-largest shareholder.²¹ In May 2026, a wholly owned Huaqin subsidiary acquired an additional 5.00% stake—100.38 million shares at ¥26.41 per share, totaling approximately ¥2.651 billion—raising Huaqin's total holding to 11.00%.²²

Market participants interpret this ownership transition through two lenses. From a strategic perspective, Huaqin's entry brings a major contract manufacturer of smartphones, tablets, and laptops directly onto the shareholder roster, potentially opening new customer channels for display drivers and power management chips. Conversely, the exit of Powerchip—the Taiwanese partner that supplied Nexchip's initial process architecture—means the founding technology provider reduced its equity stake by more than half within twelve months at prices between ¥19.88 and ¥26.41 per share. Given Powerchip's deep familiarity with the manufacturing assets, this rapid divestment provides a notable market signal.

The Wanxin structure and the "equity in name, debt in substance" question

The complexity of corporate governance is most evident in subsidiary financing structures.

Nexchip's third-phase manufacturing expansion is held within subsidiary 皖芯集成 Wanxin Integrated, established in December 2022 as a wholly owned entity. Outside investors—including state-backed financial platforms such as 农银金融资产投资 ABC Financial Asset Investment—subsequently injected capital, raising a cumulative ¥9.55 billion for equipment procurement, debt refinancing, and automotive silicon development.²³ In October 2025, Hefei Construction Investment contributed an additional ¥3.0 billion in cash. Nexchip waived its pre-emptive rights, reducing its direct ownership in Wanxin to 33.75%, while Hefei Construction Investment held 22.85%. Nexchip retained operational control and continued to consolidate Wanxin on its balance sheet.¹⁵

Financial analysts in mainland China have highlighted risks associated with these co-investment structures, often described as 名股实债 ("equity in name, debt in substance"). During the earlier second-phase project, institutional financial investors exited their holdings after approximately twenty months with a cumulative return of 13.33%—representing an annualized yield of roughly 8.42%. That fixed return pattern closely mirrors debt financing rather than equity participation in fab operating cash flows.²³ This arrangement creates an accounting contrast: reported minority profit shares declined into negative territory during downcycles even as institutional co-investors realized structured returns.²³

By the end of 2025, minority interest on Nexchip's balance sheet reached ¥6.32 billion, compared with parent-attributable equity of ¥21.77 billion.[^7] A minority equity block of that scale, combined with debt-like return expectations, directly impacts effective financial leverage, widens the gap between consolidated and attributable net income, and creates eventual cash requirements if subsidiary holdings are repurchased. While fully disclosed in statutory filings, these obligations are omitted from headline attributable earnings metrics.

The activist stress test

An institutional evaluation of Nexchip's capital efficiency highlights four key structural friction points.

Capital returns. For financial year 2024, the company proposed a token cash dividend of ¥1.00 per ten shares—equivalent to ten fen per share.²⁴ For 2025, despite record attributable net profit, management recommended no profit distribution, omitting cash dividends, bonus shares, and reserve capitalization.²⁴ Consequently, two consecutive years of profit growth yielded minimal direct cash return for public shareholders, as cash flows were directed into facility expansion.

Capacity expansion. In January 2026, Nexchip launched its fourth-phase expansion project—a ¥35.5 billion investment in a twelve-inch manufacturing line with planned monthly capacity of 55,000 wafers. The facility targets 40-nanometer and 28-nanometer image sensors, OLED drivers, and logic chips, with equipment installation scheduled for the fourth quarter of 2026 and full production targeted for the second quarter of 2028.²⁵¹³ This single capital commitment exceeds three times total 2025 revenue, expanding capacity directly into a mature-node market already facing industry-wide oversupply concerns.

Equity financing. On July 10, 2026, Nexchip completed an H-share listing in Hong Kong, raising HK$6.98 billion, with 53.6% of net proceeds designated for research and development on a 22-nanometer process platform and remaining funds allocated to AI-enabled manufacturing enhancements.¹ Twenty cornerstone investors subscribed for 107.9 million shares—representing 49.92% of the global offering—including Chipone, 奇瑞汽车 Chery Automobile, 歌尔股份 GoerTek, 泰康人寿 Taikang Life, Hillhouse's HHLR Fund, and 高毅 Gao Yi Asset Management.² Anchored by key customers and industrial partners, this cornerstone backing provides strategic validation while highlighting that the shareholder registry remains heavily aligned with corporate and industrial interests.

Valuation spread. The H-share offer price of HK$32.30 represented a substantial discount to Nexchip's A shares, which traded near ¥54 during the same week at a total market capitalization of approximately ¥120 billion.²⁶ While dual-listed Chinese firms typically trade at an A-share premium, a valuation spread of this magnitude reflects a distinct market pricing gap between international and domestic investors evaluating the same underlying corporate cash flows.

In summary, Nexchip operates with relatively transparent financial disclosures and responsive investor communications regarding pricing and factory utilization. The primary issue for public investors is alignment. Nexchip is controlled by a municipal government executing a regional industrial strategy, funded through multi-tiered equity issuance, where capital is reinvested into manufacturing scale rather than distributed as shareholder cash flows. While this structure functions effectively for municipal economic planning, it operates on a different investment logic than standard equity models prioritized by international capital markets.

The ultimate test of this corporate strategy depends on whether new manufacturing capacity can be converted into higher-margin products beyond legacy 110-nanometer LCD display drivers.

↑ Back to Top

VII. Segment Deep-Dive & The Node Transition Pivot

Understanding Nexchip's product portfolio requires looking inside a standard smartphone.

Behind the glass panel lies an array of pixels. A display driver IC takes the digital image computed by the main processor and translates it into thousands of precisely timed analog voltages—one for each column of pixels—that prompt liquid crystals to twist or organic diodes to emit light at specific brightness levels. It is fundamentally a high-voltage analog chip. Rather than requiring bleeding-edge speed or microscopic transistor sizes, it demands electrical precision and low manufacturing costs. Consequently, display drivers reside on mature process nodes: shrinking transistors offers minimal performance benefit because the chip manages voltage rather than raw computation.

Facing outward through the same glass is a CMOS image sensor—a grid of light-sensitive wells that converts photons into electrical charge and then into digital data. Image sensors present higher manufacturing hurdles than display drivers: the pixel array must remain optically clean, the underlying transistor layer is frequently stacked on a second bonded wafer, and defect tolerances are far stricter. This higher technical complexity commands higher pricing per wafer, making image sensors a primary target for a display-driver foundry seeking margin expansion.

A power management IC functions as a device's electrical distribution panel, stepping down battery voltage into the distinct, tightly regulated supply levels required by internal components. It relies on a specialized manufacturing process—BCD, which integrates bipolar, CMOS, and DMOS transistor structures on a single die—to handle both delicate logic circuits and high current loads.

Finally, logic encompasses the remaining component spectrum: timing controllers, image signal processors, connectivity chips, and compact systems-on-chip that populate the broad middle market of the semiconductor industry, where vendor competition is most intense.

Where the revenue actually comes from

In 2025, main-business revenue reflected this expanding mix: display drivers generated 58.06%, image sensors reached 22.64%, power management accounted for 12.16%, logic represented 3.87%, and microcontrollers made up 2.82%.⁴

The trajectory highlights a clear shift in product mix. Display drivers accounted for 71.2% of revenue in 2022 and climbed to 84.8% in 2023 as newer segments remained small and display driver volumes held steady. That share subsequently fell to 67.5% in 2024, roughly 60.6% in the first half of 2025, and 58.06% for the full year.¹²⁴ Concurrently, image sensors gained 5.38 percentage points of total product mix in 2025 alone.²⁷ This shift demonstrates measurable product diversification over a three-year window rather than an unfulfilled corporate target.

However, a second metric provides critical context. Revenue from 55-nanometer and finer process nodes represented 9.9% of total sales in 2024 and 10.3% in the first half of 2025.¹² Roughly nine-tenths of Nexchip's output remains on 90-nanometer and coarser manufacturing processes—precisely the mature nodes subject to industry-wide capacity expansion and heavy price competition. The company has diversified its products significantly faster than it has upgraded its underlying process nodes.

The OLED battleground

The primary strategic target in display drivers is the OLED market. As smartphone displays transition from LCD to OLED panels, driver chips require smaller process geometries—specifically 55-nanometer and 40-nanometer nodes—to integrate extensive digital compensation logic alongside high-voltage output stages, correcting for uneven emitter degradation over time. Combining dense logic with high-voltage circuitry on a single die presents technical challenges historically dominated by Taiwanese and South Korean foundries.

Nexchip has achieved initial manufacturing milestones in this segment. According to its 2025 annual report, 40-nanometer high-voltage OLED drivers, 55-nanometer image sensors, and 55-nanometer logic chips had entered mass production, while its 28-nanometer logic platform completed process development.⁴ Management publicly stated that it successfully validated a television display panel on its 28-nanometer logic platform—confirming engineering support for timing controllers, image signal processors, systems-on-chip, Wi-Fi, and codec designs—with risk production scheduled for late 2025.²⁷

Certification is not commercialisation

Converting technical achievements into commercial revenue remains a distinct operational challenge. Nexchip's record in translating process milestones into high-volume commercial production shows a mixed pattern, as documented in corporate filings.

Company project disclosures and use-of-proceeds filings highlight several timeline adjustments. First, the 90/55-nanometer image sensor platform was initially scheduled for completion in the fourth quarter of 2023, deferred to late 2024, and postponed again in January 2025 to the end of 2025.¹² Second, the 40-nanometer logic platform target moved from mid-2024 to mid-2025.¹² Third, in April 2025, Nexchip terminated its 55-nanometer and 40-nanometer microcontroller platform project outright, reallocating unspent capital toward 28-nanometer logic and OLED development.¹²

The microcontroller cancellation provides a clear case study in corporate resource allocation. Microcontrollers featured prominently in the company's initial IPO diversification narrative and received dedicated funding. However, the program was discontinued before reaching commercial scale, with capital shifted to smaller process nodes. For public investors evaluating the 28-nanometer roadmap, this history illustrates that technical development does not automatically translate into profitable manufacturing volume, given a three-year record that includes one project cancellation and multiple schedule adjustments.

Reallocating capital from underperforming projects reflects disciplined portfolio management, and the shift was formally disclosed in statutory filings. Nevertheless, these developments indicate that the node transition remains an ongoing operational test. While Nexchip demonstrated its capacity to license, transfer, and ramp mature high-voltage processes to high volumes, it has yet to prove it can independently develop advanced-mature platforms and capture profitable market share against established competitors such as 中芯国际 SMIC, 华虹半导体 Hua Hong, and Taiwanese foundries that have operated 40-nanometer and 28-nanometer lines for over a decade.

The automotive option, and how to weigh it

A fourth element of the diversification strategy involves automotive semiconductors. Financing disclosures for the third-phase expansion identify it as an automotive chip project, while the fourth-phase expansion plan explicitly targets automotive-adjacent image sensors and logic.²³²⁵ The automotive sector represents a strategic market for mature-node foundries: modern vehicles require large volumes of power management ICs, image sensors, and display controllers. Furthermore, automotive qualification processes are rigorous and time-consuming, creating high customer switching costs.

Achieving automotive-grade qualification requires extensive reliability testing over multiple years. Once a chip design is qualified at a specific manufacturing facility, automotive customers rarely switch foundries due to requalification costs and supply chain risks. Successfully converting production capacity to automotive-grade components would structurally strengthen Nexchip's customer retention compared to consumer electronics segments.

However, qualification does not immediately yield commercial revenue. Automotive design wins involve multi-year lead times—often three to five years from initial qualification to full-volume production—and purchasing decisions remain controlled by major tier-one suppliers with strict multi-sourcing requirements. While the participation of 奇瑞汽车 Chery Automobile as a cornerstone investor in Nexchip's Hong Kong listing signals commercial alignment,² it does not constitute a binding order book. Consequently, the automotive segment represents potential long-term optionality rather than an immediate driver of earnings.

The live test: 2026

Financial performance in 2025 and mid-2026 illustrates the underlying economic tension facing the business.

Operational results in 2025 initially reflected strong factory loading. Capacity utilization reached 100.8% in the first half of 2025 as operational refinements allowed production to exceed nominal capacity.¹² First-half revenue rose 18.21% to ¥5.198 billion, net profit attributable to shareholders increased 77.61% to ¥332 million, and gross margin expanded to 25.76%.²⁸ By the end of the third quarter, cumulative revenue reached ¥8.13 billion with a gross margin of 25.90%, outperforming the mature-node industry average of approximately 22%.²⁹ For the full year, research and development expenditure increased 13.20% to ¥1.454 billion, representing 13.35% of total revenue—a substantial investment rate for a mature-node foundry.⁴

However, financial results weakened in early 2026. In the first quarter of 2026, revenue grew 13.41% year-on-year to ¥2.912 billion, but net profit fell 62.61% to ¥50.66 million, while net profit excluding non-recurring items declined 68.38%.³⁰³¹ Gross margin for the quarter dropped to 21.26%.³² Management attributed the profit decline to intensified market competition that depressed average selling prices, higher fixed-asset depreciation charges, and fair-value losses on equity investments in listed companies.³² In the first half of 2026, revenue increased 12.40% to ¥5.766 billion, even as attributable net profit fell 26.12% to ¥245 million.⁷

This divergence between revenue and profit highlights the structural dynamics of pure-play foundry operations. Although manufacturing lines remain fully loaded and shipment volumes continue to expand, newly added capacity generates substantial depreciation overhead while market-wide price competition reduces per-wafer realizations faster than volume increases. Throughout its operational history, Nexchip's top-line growth has reflected its manufacturing capacity expansion, while its profitability has remained bound to external market pricing power.

↑ Back to Top

VIII. Playbook: Business & Investing Lessons

Every good business story leaves behind a set of transferable lessons. Nexchip's are unusually clean because the company ran the full experiment in public view in under a decade.

Lesson one: in a commodity manufacturing business, volume share is a fact, not a moat. Nexchip achieved something extraordinarily rare: taking the global top position in a product category from an entrenched incumbency, starting from scratch, in a country with no prior presence in that segment. Yet when the cycle turned, market leadership provided virtually no pricing power. The reason is structural rather than managerial: buyers are fabless design houses whose operating model centers on allocating wafer volume across multiple foundries. Their products, at 90-nanometer and 110-nanometer nodes, can be transferred between foundries at a cost measured in mask sets and months rather than years or architectural redesigns. Market share protects a firm from being ignored; it does not protect it from being repriced. The critical investment question is never how much market share a company holds, but what customers do when a competitor offers a 15% discount.

Lesson two: operating leverage is a two-sided instrument, and the second side arrives without warning. A twelve-inch fab operates substantially as a fixed-cost asset. Nexchip's gross margin swung from negative territory to roughly 43% and back to approximately 20% within four years, on a process technology and customer base that barely changed.[^7]¹² The key indicator is not gross margin alone, but the ratio of depreciation to revenue. When depreciation approaches 40% of sales, small movements in average selling prices produce outsized swings in profit. Modeling a capital-intensive manufacturer requires building the depreciation schedule first and the revenue forecast second, because depreciation costs are fixed while revenue is subject to market demand.

Lesson three: municipal venture capitalism solves the hardest problem in capital-intensive industry, and creates a different one. No private investor would have funded four consecutive years of losses on the scale Nexchip absorbed between 2018 and 2020, when the company burned more than two yuan for every yuan of revenue in its early years.[^7] Hefei's state capital filled that gap because its return was evaluated in industrial cluster formation, local employment, and supply-chain security rather than financial internal rates of return. Patient state capital was a necessary condition for initial scale. However, listing on public exchanges introduced a second constituency with different financial objectives. Those priorities diverge at pivotal capital allocation decisions: returning cash to shareholders versus building additional fab capacity. To date, management has consistently prioritized facility expansion—including recommending no profit distribution following record attributable earnings in 2025.²⁴ Investors in state-anchored industrial champions are not purchasing an immediate claim on free cash flow; they are investing in long-term manufacturing scale on the controlling shareholder's timetable.

Lesson four: the node treadmill is funded from the business the treadmill is meant to escape. This structural reality represents a core challenge of Nexchip's position and applies broadly to technology firms advancing up the process capability curve. Capital needed to develop 40-nanometer and 28-nanometer process nodes must be generated by selling 90-nanometer and 110-nanometer wafers—the exact legacy segments facing severe price erosion. A company must run faster on legacy nodes to fund its exit from them, while incumbents already operating at smaller nodes enjoy superior unit economics and can price aggressively in target growth segments. Specialized financing cannot remove this dynamic; success depends on execution speed, making schedule delays costly.

A related principle applies to evaluating companies undergoing technological transitions: the schedule is the strategy. A process platform delivered on schedule enters a market before incumbent pricing fully adjusts. A platform delayed by two years enters an environment already commoditized by earlier entrants. This context highlights why historical platform delays carry structural financial weight. Each schedule slip is not merely a deferred revenue milestone, but a permanent reduction in the peak pricing the eventual product can command.

Lesson five, implicit in all of the above: read the second-order disclosure. The most critical operational insights in this narrative were not located in headline financial metrics. They emerged from the gap between consolidated and attributable profit, the waiver of pre-emptive rights during subsidiary capital raises, the structured returns realized by early financial co-investors, block equity trades by founding partners, and project cancellation notices embedded in statutory use-of-proceeds filings. Headline revenue and earnings figures describe industry cycles; footnote disclosures reveal the shareholder's actual economic position.

Taken together, these lessons describe an enterprise that executed the capital-intensive task of building advanced domestic manufacturing capability, only to find that markets reward capability primarily when it is scarce. This outcome reflects the structural reality of commodity manufacturing rather than a flaw in operational execution—a condition that investors in the sector must underwrite explicitly.

With the underlying business mechanics established, the final question is competitive: where does Nexchip sit in the global semiconductor power structure?

↑ Back to Top

IX. Strategic Position, 7 Powers & Porter's 5 Forces

Hamilton Helmer's 7 Powers, applied honestly

Scale economies — real but relative, and the relativity is brutal. Nexchip's twelve-inch manufacturing strategy lowered per-wafer costs compared to legacy eight-inch lines, and expanding monthly capacity beyond 100,000 wafers provided meaningful fixed-cost absorption within its core niche.¹³ However, scale power is measured relative to competitors rather than historical performance. In the first quarter of 2026, Nexchip's quarterly revenue of roughly US$400 million represented just 0.8% of the global foundry market, compared to TSMC's 72% market share.⁵ Within display drivers specifically, the pattern reverses: Nexchip led the sector, holding roughly one-quarter of global display driver foundry volume in 2024.⁶ Consequently, Nexchip possesses category-level scale power but lacks industry-wide scale authority. It can out-cost smaller display driver rivals, but it cannot match the investment scale of broader foundries such as SMIC, Hua Hong, or TSMC across wider platform portfolios.

Process power — the strongest of its powers, and still bounded. A decade of operating high-voltage manufacturing lines at scale generated accumulated engineering expertise, including process recipe libraries optimized for driver-chip color depth and energy efficiency, yield refinements gained from processing millions of wafers, and proprietary touch-and-display integration platforms built on top of its licensed baseline.⁹ This operational experience helps explain why Nexchip's gross margin ran modestly above the mature-node industry average during the sector recovery.²⁹ Nevertheless, this advantage remains bounded because high-voltage manufacturing processes are licensable, replicable, and increasingly commoditized, and because the company's core capabilities remain concentrated in node ranges with severe industry-wide overcapacity.

Cornered resource — weak, and frequently overstated. The potential candidates for a cornered resource include Hefei's municipal subsidies, infrastructure support, industrial cluster proximity, and Powerchip's founding technology transfers. While each offers operational benefits, none is exclusive. Municipal governments across China offer similar incentive packages to local foundries, driving widespread mature-node capacity expansion. Furthermore, the technology heritage represents a diminishing connection: founding partner Powerchip reduced its equity holding from 19.08% to roughly 8% across two block transactions within twelve months.²¹²² An asset that competitors can secure on comparable terms provides a temporary cost advantage rather than a cornered resource.

Switching costs — low at the base, moderate at the frontier. For standard 90-nanometer LCD drivers, customer switching costs involve a mask set and a standard qualification cycle. For 40-nanometer OLED drivers with custom high-voltage integration or automotive-qualified image sensors with strict requalification requirements, switching barriers increase substantially. This dynamic underscores the strategic intent of Nexchip's node transition: the company is targeting product categories that offer stronger customer retention. Achieving this shift at scale remains an open operational challenge, as chips produced at 55 nanometers or finer accounted for roughly one-tenth of total revenue in recent reporting periods.¹²

Counter-positioning — absent, and worth stating plainly. Nexchip operates a standard pure-play foundry model without structural business-model asymmetries that incumbents cannot copy. Its competitive attributes center on manufacturing cost and category focus rather than counter-positioning. Similarly, branding and network economies offer minimal competitive differentiation in business-to-business wafer manufacturing, where contracts are negotiated on price, yield, and delivery schedules.

Porter's Five Forces: the war-game

Rivalry — extremely high, and structurally so. The 28-to-110-nanometer node spectrum represents the primary intersection of mainland Chinese capacity expansion, established Taiwanese foundries, and broader industry overcapacity. Foundries including SMIC, Hua Hong, UMC, VIS, and PSMC compete for overlapping order volumes, with mainland Chinese foundries actively winning market share from GlobalFoundries, PSMC, and Samsung through aggressive wafer pricing.³³ In a capital-intensive sector with high fixed costs and low marginal production expenses, intense price competition during cyclical downturns represents an ongoing structural condition rather than a temporary disruption.

Bargaining power of suppliers — very high, with a geopolitical multiplier. Semiconductor manufacturing equipment remains concentrated among a small group of global vendors—including ASML, Applied Materials, Tokyo Electron, and Lam Research, alongside domestic suppliers such as Naura Northern Microelectronics—giving equipment makers substantial pricing power and lead-time control. International export controls add further operational complexity: restrictions on advanced manufacturing tools have expanded, and while mature-node equipment remains largely accessible, regulatory thresholds remain subject to policy shifts. Nexchip's planned R&D investment into a 22-nanometer process platform, funded through its Hong Kong H-share listing proceeds, moves its technology roadmap closer to those regulatory boundaries.¹

Bargaining power of buyers — high, and demonstrated rather than theoretical. Buyer power is reflected in customer concentration metrics and standardized chip designs: historically, Nexchip's top five fabless customers generated over 60% of total revenue, with a single design house contributing the majority of sales during early growth periods.¹² Customer concentration has gradually declined as the product mix expanded, with the largest customer's revenue share falling from roughly 61% to approximately 31% across reported periods.¹² While customer diversification mitigates single-buyer reliance, fabless design houses retain significant leverage to reallocate wafer orders across alternative foundries based on pricing terms.

Threat of new entrants — low in the conventional sense, but structurally misleading. Building a greenfield twelve-inch semiconductor fab requires billions of dollars in capital expenditure and several years of construction, presenting a formidable barrier to private venture capital. However, those capital requirements do not deter municipal local-government vehicles operating under regional economic development mandates. In mainland China's mature-node foundry landscape, the primary entrant risk stems not from commercial startups, but from competing municipal investments replicating similar industrial development frameworks.

Threat of substitutes — low at the wafer level, high at the product level. While no direct technological substitute exists for silicon wafer manufacturing, product-level substitution risk is significant. The industry transition from LCD to OLED displays represents a major shift within Nexchip's primary product category, requiring ongoing capital investment in smaller process nodes to maintain manufacturing relevance. The principal risk is not the obsolescence of silicon manufacturing, but the risk that customer demand shifts away from the specific process nodes a fab is equipped to produce.

The regulatory overhang

In December 2024, the Office of the United States Trade Representative initiated a Section 301 investigation into China's acts, policies, and practices concerning the legacy semiconductor sector, targeting mature-node chips defined as 22 or 28 nanometers and larger across defense, automotive, medical device, aerospace, and communications applications.³⁴ Nexchip's manufacturing portfolio falls squarely within this operational definition. Although the investigation does not cite the company specifically, potential trade measures—such as tariffs, domestic content requirements, or procurement restrictions on Chinese mature-node silicon—represent an external policy risk that falls directly on its operating model.

Taken together, these competitive dynamics describe an industry characterized by high barriers to entry alongside low barriers to intra-industry competition—a challenging combination for long-term return on capital, requiring massive initial investment while providing limited pricing protection against cyclical market competition.

↑ Back to Top

X. Bear vs. Bull Case & Key Investor KPIs

Myth versus reality

Before weighing the investment cases, three consensus narratives deserve scrutiny against the empirical record, as each appears regularly in coverage of the company and each proves flawed.

Myth: Nexchip is a beneficiary of the AI boom. The operational reality is far more modest. TrendForce attributed the company's rise to eighth in global foundry rankings in the first quarter of 2026 to pull-in orders for television and personal computer peripheral chips—a temporary restocking of consumer electronics rather than demand for artificial intelligence accelerators.⁵ The connection to AI is indirect at best: AI-capable end devices require display drivers, image sensors, and power management circuits just like standard hardware. While a portion of Nexchip's Hong Kong listing proceeds was designated for AI-enabled production capabilities, that capital funds automated manufacturing systems inside its fabs rather than the production of AI silicon.¹ Investors purchasing the stock as an AI pure-play are simply buying into a standard consumer electronics cycle under a different label.

Myth: State backing protects downside risk. Municipal state support has reliably shielded the enterprise itself, absorbing years of operating losses that commercial capital would have rejected. However, it has not protected the stock price or delivered returns for minority shareholders. Nexchip's share price closed 2023 below its initial public offering price, and despite record attributable net profit in 2025, management recommended no cash dividend or profit distribution.¹⁴²⁴ Sovereign investment horizons and public equity returns operate on fundamentally different logics; patient municipal capital does not guarantee minority shareholder gains.

Myth: Display drivers are a declining legacy asset the company is abandoning. The business is neither contracting nor easily abandoned. Display drivers continued to generate the majority of corporate revenue in 2025, and the industry-wide transition from LCD to OLED represents a product upgrade within the segment rather than an exit from it. OLED drivers represent higher-value components requiring finer process nodes, serving as the primary target for Nexchip's advanced manufacturing development.⁴ The accurate strategic framing is not diversification away from display drivers, but diversification around a core display franchise that is moving up the process ladder.

The bull case

One: Product diversification is measurable and expanding. This argument rests on audited disclosures rather than executive projections. Display drivers declined from 84.8% of revenue in 2023 to 58.06% in 2025 while total sales expanded, demonstrating that non-display segments grew significantly faster than the core business.¹²⁴ CMOS image sensors alone expanded to 22.64% of total revenue in 2025.⁴ If image sensors, power management, and logic continue to gain several percentage points of product mix annually, Nexchip will transition into a diversified mature-node foundry with a more stable earnings profile across semiconductor cycles.

Two: Domestic localization tailwinds provide policy-backed demand. Chinese fabless chip designers face sustained commercial and regulatory pressure to qualify domestic manufacturing capacity, and Nexchip remains one of only three mainland pure-play foundries operating at commercial scale. The acquisition of an 11.00% equity stake by contract manufacturer Huaqin and the cornerstone participation of chip designer Chipone in the Hong Kong IPO demonstrate this structural alignment: downstream customers and contract manufacturers taking equity positions to secure factory capacity.²²² This domestic alignment creates a captive demand channel that foreign foundry competitors cannot replicate.

Three: Cyclical demand recovery coincides with a moderating depreciation burden. TrendForce reported improving factory utilization rates across pure-play foundries in early 2026, with mature process pricing reaching a cyclical trough and foundries signaling potential wafer price increases in the second half of the year.⁵ TrendForce attributed Nexchip's rise to eighth place globally to restocking orders for television and personal computer peripheral chips.⁵ Concurrently, semiconductor manufacturing equipment on twelve-inch lines typically depreciates over five to seven years, meaning initial production tools installed from 2017 onward are nearing the end of their schedule. Firming wafer prices combined with a stabilizing depreciation base create the specific financial leverage required to expand operating margins.

The bear case

One: Mature-node overcapacity represents a persistent structural challenge. Mainland Chinese foundries continue to commission twelve-inch mature-node capacity simultaneously, with Nexchip adding 55,000 wafers per month in its fourth-phase expansion.²⁵ While individual foundries act logically by building scale, collective expansion caps potential price recoveries and intensifies cyclical downturns. Results from the first half of 2026 illustrate this dynamic: sales increased 12.40% while attributable net profit fell 26.12%, demonstrating how capacity expansion without pricing power depresses earnings even during periods of volume growth.⁷

Two: Diversification has expanded across products rather than manufacturing nodes. Approximately nine-tenths of corporate revenue remains tied to process nodes of 90 nanometers and coarser—the most competitive segment of the foundry market. The advanced platforms intended to lift realization prices have experienced recurring schedule adjustments, including two timeline postponements for the image sensor platform, a delayed launch for 40-nanometer logic, and the complete cancellation of the microcontroller program.¹² Development programs for 28-nanometer and 22-nanometer process nodes represent larger, higher-risk iterations of this same technological transition.

Three: Sustained capital expenditure consumes cash flow and dilutes equity holders. Free cash flow remained deeply negative across every reporting period following Nexchip's public listing, and record attributable profit in 2025 resulted in a proposed nil dividend.[^7]²⁴ Corporate expansion has relied on external financing, including its A-share initial public offering, a Hong Kong H-share listing, ¥9.55 billion in private equity raised for subsidiary Wanxin Integrated, a ¥3.0 billion direct injection by controlling shareholder Hefei Construction Investment, and substantial debt financing.^1412315 This multi-tiered capital structure dilutes minority equity holders at both parent and subsidiary levels, while widening the divergence between consolidated net profit and parent-attributable net income.

Four: Executive shifts and founding partner divestments signal operational friction. The reliance on a 72-year-old chairman serving as acting general manager during a complex process migration, a 20% reduction in designated core technical personnel, and the sale of over half of founding partner Powerchip's equity holding within twelve months present cumulative governance concerns.^18192122 Combined, these developments suggest caution regarding long-term execution timelines.

The three KPIs that actually matter

Everything above compresses into three critical operational metrics.

One: Gross margin evaluated alongside capacity utilization. Evaluating gross margin or factory loading in isolation is insufficient; the two metrics must be analyzed together. Utilization measures production volume, while gross margin at full capacity reveals per-wafer realization prices. Nexchip operated near full capacity in 2024 while generating a 25% gross margin, compared to a 43% gross margin achieved at an 85.9% utilization rate in 2022.¹² The thesis that Nexchip possesses a structural cost advantage depends on whether gross margins can hold in the upper twenty-percent range during periods of full factory loading. If utilization remains high while gross margins decline toward 20%, industry-wide pricing pressure will outweigh volume growth.

Two: Revenue contribution from 55-nanometer and finer process nodes. This percentage serves as the primary benchmark for determining whether process engineering advances are translating into commercial sales. Revenue from nodes of 55 nanometers and finer represented 9.9% of sales in 2024 and 10.3% in the first half of 2025.¹² Because advanced-mature nodes command higher wafer prices and create stronger customer switching barriers, this metric measures both margin expansion potential and competitive differentiation. An increase in this share toward 20% or higher would validate the product pivot, whereas a flat metric near 10% following substantial R&D expenditure on 28-nanometer and 22-nanometer platforms would indicate execution delays.

Three: Non-display driver revenue share, specifically CMOS image sensors. Display drivers generated 58.06% of corporate revenue in 2025.⁴ Expanding the non-display driver share above 40%—anchored by image sensors, which require complex processing steps and command higher per-wafer realizations—is essential to transforming Nexchip from a single-product manufacturer into a diversified foundry. Three consecutive years of product diversification represent the most tangible operational progress in the company's public record.

Where that leaves the "why win / why not" spine

Nexchip's long-term commercial thesis succeeds only if three conditions align: mature-node wafer pricing stabilizes; advanced-mature process platforms convert into high-volume commercial revenue on schedule; and fourth-phase production capacity ramps into a recovering market rather than an industry glut. If those conditions are met, category leadership, established high-voltage process engineering, domestic localization policies, and a depreciating legacy asset base could support operating returns above recent historical averages.

Conversely, the thesis fails if mature-node pricing competition persists through the next expansion phase. Under that scenario, depreciation from third- and fourth-phase facilities will weigh on an income statement characterized by expanding wafer volumes and declining unit realizations—mirroring performance in the first half of 2026—turning the equity into a speculative instrument dependent on cyclical timing and sustained capital dilution.

Empirical evidence rejects the simplest optimistic claim: that global volume leadership in display driver chips creates a protective competitive moat. Volume scale failed to shield profitability during the 2023 downcycle and did not prevent margin contraction in 2026. What remains to be proven is whether a decade of high-volume manufacturing experience, patient municipal backing, and proximity to mainland panel and hardware manufacturing can be converted into a defensible position at advanced process nodes.

Eleven years after Powerchip transferred its initial process blueprints to a municipal investment vehicle in inland Anhui, that strategic question remains open. The manufacturing facilities were constructed, and global market share was secured. The underlying profit margin remains subject to ongoing market negotiation, one wafer at a time.

References

1. China's Nexchip to raise $890 million in Hong Kong share sale — Reuters via Yahoo Finance, 2026-07-08 ↩↩↩↩↩

2. 晶合集成(02249)今起招股，引入集创、奇瑞、高毅、高瓴、常春藤等20家基石，7月10日香港上市 — 腾讯新闻, 2026-06-30 ↩↩↩↩

3. 晶合集成：2025年净利润7.04亿元 同比增长32.16% — 证券时报网, 2026-03-27 ↩↩↩

4. 晶合集成2025年净利润同比增长32.16% 产能利用率维持高位 — 证券时报网, 2026-03-27 ↩↩↩↩↩↩↩↩↩↩

5. Strong AI Demand and Early Consumer Electronics Inventory Build Drive Top 10 Foundries to 3.7% QoQ Revenue Growth in 1Q26 — TrendForce, 2026-06-12 ↩↩↩↩↩↩

6. NexChip and Samsung Dominate Display Driver IC Market — Display Daily ↩↩↩↩

7. 晶合集成：上半年净利润2.45亿元，同比下降26.12% — 界面新闻, 2026-08 ↩↩↩

8. 晶合集成科创板IPO获受理 募资120亿用于12英寸晶圆制造 合肥国资现身股东名单 — 财联社 ↩↩

9. 【报告】晶合集成：从合肥走向香港，一家大陆晶圆代工厂的资本出海 — 腾讯新闻, 2026-04-03 ↩↩↩↩↩↩↩↩↩

10. 国内12英寸晶圆代工产能第三大企业，合肥晶合集成IPO成功 — 电子工程专辑, 2023-05-06 ↩

11. 晶合集成管理层现人事变动：总经理辞职 72岁董事长蔡国智将代行职权 — 财联社, 2025-02-14 ↩

12. 晶合集成赴港上市：60%营收依赖单一产品，核心项目两度延期 — 财中社 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

13. 企业简介 — 合肥晶合集成电路股份有限公司 ↩↩↩↩

14. 2023年IPO首发募资额下滑近四成，晶圆厂包揽年度前三 — 澎湃新闻, 2023-12-29 ↩↩↩↩↩

15. 合肥建投30亿元增资皖芯集成，晶合集成放弃优先认购权，但仍有控制权 — 每日经济新闻, 2025-10-16 ↩↩↩

16. 晶合集成前三季度归母净利预增超7倍 28nm OLED驱动芯片拟明年量产 — 财联社, 2024-10-09 ↩↩

17. 晶合集成：2025年净利润4.67亿元，同比下降3.19% — 证券之星, 2026-02-27 ↩

18. 晶合集成总经理离职 蔡国智代行职权 — 证券时报网, 2025-02-14 ↩↩↩

19. 晶合集成总经理及核心技术人员蔡辉嘉离任 — 新浪财经, 2025-02-14 ↩↩↩↩

20. 晶合集成(688249)公司高管 — 新浪财经 ↩

21. 华勤技术23.93亿元入股晶合集成，成为第四大股东 — 电子工程专辑, 2025-07-31 ↩↩↩

22. 26.51亿，华勤技术完成晶合集成5%股权受让 — LED在线, 2026-05-21 ↩↩↩↩

23. 晶合集成再募近百亿增资三厂，名股实债争议再起 — 乐居财经 ↩↩↩↩↩

24. 晶合集成（688249）2025年年报简析：营收净利润同比双双增长 — 搜狐财经, 2026-03-28 ↩↩↩↩↩

25. 半导体扩产：晶合355亿四期项目开建、华虹官宣82亿收购 — 国际电子商情 ↩↩↩

26. 每周股票复盘：晶合集成（688249）H股挂牌上市 — 凤凰网财经, 2026-07 ↩

27. 晶合集成：全球显示驱动晶圆代工龙头以28纳米"点亮TV"为锚 — 腾讯新闻, 2026-04-03 ↩↩

28. 晶合集成上半年营收51.98亿元同比增18.21%，归母净利润3.32亿元同比增77.61% — 新浪财经, 2025-08-28 ↩

29. 晶合集成的前世今生：2025年Q3营收81.3亿领先同业，毛利率25.9%高于行业平均3.76个百分点 — 新浪财经, 2025-11-01 ↩↩

30. 晶合集成：2026年一季度净利润5065.86万元 — 东方财富网, 2026-04-24 ↩

31. 晶合集成2026年一季报解读：扣非净利大降68.38% 经营现金流增64.36% — 新浪财经, 2026-04-23 ↩

32. 晶合集成：2026年一季度净利润5065.86万元 — 搜狐财经, 2026-04-24 ↩↩

33. TSMC Gains Foundry Share But Others Lose Ground — Mark LaPedus, Semiconductor Analysis ↩

34. Section 301 and China: Mature-Node Semiconductors — Congressional Research Service, IF12958 ↩