Shenzhou International: 22.6% H1 Gross Margin and 75.9% Customer Concentration Leave No Conservative Margin of Safety at HK$36.58

The labor accounts show the same squeeze. H1 wages and salaries rose 1.1% to RMB3.872 billion, retirement contributions increased 13.2% to RMB475 million and other employee benefits rose 33.9% to RMB191 million. Total employee benefit expense increased 3.3% to RMB4.538 billion while sales fell 5.3%. Shenzhou was paying for more people, higher wages and capacity that had not yet reached mature productivity as the revenue base shrank.

Currency amplified the operating problem through three separate channels. Commercial conversion came first: most exports are settled in USD, while Shenzhou reports in RMB. Management says H1 sales would have declined approximately 1.8% in USD terms rather than 5.3% in RMB terms. Applying that constant-currency decline to H1 2025 revenue implies roughly RMB14.70 billion of H1 2026 revenue before the reporting-currency hit, approximately RMB0.52 billion above the reported RMB14.18 billion. Monetary items then generated an RMB506.2 million P&L foreign-exchange expense, versus an RMB126.4 million gain a year earlier, a RMB632.6 million negative swing. Translation of overseas operations produced an RMB412.6 million loss in other comprehensive income. The exposures are economically related but distinct in the accounts.

The H1 profit bridge follows those three layers. Profit before tax fell by approximately RMB1.59 billion. About RMB0.96 billion of that deterioration occurred before the year-on-year change in P&L FX remeasurement; gross profit alone accounted for approximately RMB0.86 billion. The RMB632.6 million remeasurement swing explains almost all of the remaining pretax deterioration. Tax expense then fell from roughly RMB455 million to RMB135 million, cushioning the net-profit decline by approximately RMB320 million. The effective tax rate fell to about 6.6% from 12.5%, partly reflecting deferred tax and the geographic tax mix. The tax benefit was not an operating recovery.

The margin shock followed an earlier softening in profitability during 2025. FY2025 revenue grew 8.1% to RMB30.994 billion, but gross margin declined from 28.1% to 26.3% and attributable profit fell 6.7% to RMB5.825 billion. Management attributed the margin decline principally to higher average wages, low initial efficiency at a new Cambodian garment factory and sharing part of the increased U.S. tariff burden with customers. The cost problem was already visible when the group entered 2026. RMB appreciation and another leg of capacity expansion made it worse.

Customer concentration is the second major issue and an unusually severe one. In H1 2026, anonymous Customers A, B and C generated RMB4.752 billion, RMB3.368 billion and RMB2.642 billion respectively, equivalent to 33.5%, 23.8% and 18.6% of group revenue. Combined concentration reached 75.9%, versus 71.9% in H1 2025. The annual filings show a similar climb in the combined A/B/C percentage from 69.6% in 2022 to 69.6% in 2023, 71.1% in 2024 and 74.4% in 2025. Shenzhou had a fourth disclosed 10%-scale customer in 2022 and 2023.

Those letters are anonymous, period-specific filing labels. I do not assign A, B or C to Nike, Uniqlo, Adidas, Puma or any other brand. Public descriptions of Shenzhou frequently associate it with major global sportswear and casualwear companies, but the filing does not establish which anonymous letter corresponds to which named customer. The economic conclusion needs no such identification: three accounts now determine about three quarters of sales, and the biggest alone determines one third.

Capacity is the third issue. The second Tay Ninh fabric plant in Vietnam has reached 100 tonnes of daily capacity; overseas assets have continued to rise; Cambodia is still moving through efficiency improvement; and management plans to begin construction of an Indonesian garment factory in September 2026, with phased operation from Q4 2027. H1 2026 capital expenditure was RMB743 million, roughly half machinery and 45% factory buildings or related land/use assets, after RMB3.004 billion of FY2025 capex. Geographic diversification can lower concentration in China and give customers more sourcing options. It also commits more fixed cost before sportswear orders have demonstrably recovered.

That capacity bet frames the stock debate. Bulls see a temporary utilization trough: Shenzhou has lived through customer destocking, pandemic closures and capacity ramps before, and an eventual normalization of orders plus better utilization could restore gross margin toward the mid-20s without heroic revenue growth. Bears see structural change: customers have more sourcing options, tariffs are raising the economic cost of Asian production, wages are rising across Shenzhou's footprint, concentration has risen rather than fallen and the company is adding fixed cost before demand catches up. Both readings have genuine evidence behind them.

Peer results sharpen the disagreement. Crystal International increased H1 2026 revenue 5.4% and net profit 10.6%. Eclat Textile increased H1 revenue 3.2% and net profit 48.5%, with Q2 gross margin recovering to 30.0%. Shenzhou simultaneously posted -5.3% revenue and -40.0% profit. Huali Industrial, a footwear OEM rather than an apparel peer, also suffered badly, with H1 profit reported down 43%, showing that tariff, customer and cost pressure is not unique to Shenzhou. Yet Eclat and Crystal prove that the whole manufacturing industry did not mechanically suffer a 40% earnings collapse. Shenzhou's own utilization, currency mix, customer mix and expansion timing matter.

The balance sheet buys Shenzhou time. At June 2026, deposit-type financial assets including cash, time deposits and fair-value financial assets totaled about RMB28.10 billion against RMB14.76 billion of interest-bearing bank borrowing, leaving roughly RMB13.33 billion on this broader net-deposit definition. The loans are partly deliberate natural hedges: the company mainly earns USD and reports procurement largely in RMB, so it carries USD and HKD loans to hedge future USD cash flows. At June, approximately RMB2.16 billion of borrowings were USD-denominated and RMB6.32 billion were HKD-denominated. Calling this simply “net cash” misses both the liquidity strength and the hedge structure.

Inventory deserves scrutiny, but not an automatic bearish reading. It rose 9.0% from December to RMB7.317 billion while H1 revenue declined. Using average December-to-June inventory against H1 cost of sales gives an indicative inventory cycle of approximately 116 days, versus about 109 days using average FY2025 inventory against FY2025 cost of sales. That is deterioration. I could not retrieve a reliable June 2025 inventory comparison from a primary interim report in the final source set, so the seasonal conclusion remains incomplete. The evidence points to a “potential early order signal, not yet proven channel blockage.”

The old business quality has not disappeared from the financials. Across 2022–2025, operating cash flow totaled approximately RMB20.68 billion versus roughly RMB21.19 billion of cumulative reported net income, a 0.98x cash-conversion ratio. FY2025 alone generated RMB5.55 billion of operating cash flow. That argues against an accounting-quality problem. The harder question is how much of the cash belongs to shareholders after maintaining and expanding the manufacturing base.

I therefore view Shenzhou as a company in transition. The old moat is real enough to have produced a long history of above-peer profitability and strong cash generation, but H1 2026 is a live test of whether that moat can still absorb customer bargaining power, wage inflation, tariff sharing, RMB strength and multi-country expansion without permanently resetting returns on capital. The 78/100 historical Buffett-style screen supplied in the research brief is useful only as evidence that there is something worth stress-testing. It has no bearing on the present valuation judgment.

Vertical history, financial review and capital-market narrative

Shenzhou's listed history runs from a small-market-cap manufacturer to, for now, a former quality-growth favorite trying to prove that a globalized production network will restore rather than dilute its economics. The offer price in the November 2005 Hong Kong IPO was HK$2.625 per share. HKEX substantial-shareholder records shortly after listing showed Ma Jianrong with an interest in 900 million shares, then 72.29% of the company. That established the founder-family control that has remained an important governance feature. The company itself says it has been listed in Hong Kong since November 2005.

The operating roots predate the listed Cayman holding company, with Ningbo at the group's industrial center. The durable decision was to build textile processing and garment making around each other instead of remaining a pure cut-and-sew contractor. That choice raised capital intensity and operating complexity, but brought fabric development, dyeing consistency, garment construction and customer sampling closer together. The model later became particularly valuable to sportswear customers, where technical fabric performance and repeatability matter more than in the lowest-end garment trade. Company filings describe Shenzhou as a vertically integrated knitwear manufacturer; the accounts nevertheless consolidate the operation into one reportable manufacturing segment.

The first post-listing stage tested the model at scale. Investors initially had to decide whether Shenzhou was simply another China-export garment company. The 2008 financial crisis provided a brutal stress test: the share price reached HK$0.80 on 27 October 2008. The industrial model survived and kept expanding. In hindsight, the share-price collapse was much larger than the permanent impairment to the business.

During the second stage, broadly the 2010s, the moat and offshore footprint took shape. Shenzhou increased overseas capacity in Vietnam and Cambodia while retaining high-value textile and manufacturing capabilities in China. The strategic logic was already becoming clear before today's tariff regime: wages in coastal China were rising, customers wanted supply-chain diversification, and adding Southeast Asian capacity made Shenzhou more useful to multinational brands. Historical filings describe continued lean-production and automation investment as the Vietnam and Cambodia operations grew in scale.

This period also changed how investors valued the company. Consistent earnings, high margins, cash generation and customer entrenchment turned the equity story from “Chinese apparel exporter” into “highest-quality Asian manufacturing compounder.” The multiple expanded with that change. The company's eventual May 2021 share-price peak of HK$207.60 cannot be explained by factory earnings growth alone; it also represented an unusually high valuation placed on supply-chain scarcity, execution quality and long-duration growth.

The pandemic distorted the third stage. Apparel demand first collapsed, then rebounded unevenly. Supply chains became a strategic issue for brands, initially strengthening the case for large, resilient suppliers. Shenzhou's own 2021 interim report argued that pandemic disruptions were causing brand owners to pay more attention to the regional layout and resilience of suppliers' production bases. The same multi-country network was exposed to factory shutdowns and health restrictions. In Vietnam, management estimated during 2021 that every 14 days of pandemic controls could reduce annual group capacity by roughly 1–2%.

The fourth stage brought the 2022–2023 reset. FY2022 revenue was RMB27.781 billion, while 2023 revenue declined 10.1% to RMB24.970 billion as European and U.S. sales fell sharply. Gross margin still recovered from 22.1% in 2022 to 24.3% in 2023, helped by better utilization in the second half. Net profit was essentially flat at around RMB4.56 billion in both years. Factory utilization mattered almost as much as the top line: fewer sales accompanied by better manufacturing economics could still protect earnings.

That episode provides the best evidence for today's bull case. In H1 2023 gross margin was around 22.4%; by H2 it had recovered to around 25.8% as utilization improved. Investors looking at H1 2026's 22.6% margin see a familiar number and a familiar recovery template. The comparison is incomplete, though. H1 2026 also includes a much larger FX hit, a higher labor-cost base, tariff sharing and another wave of overseas capacity.

The fifth stage was the 2024 rebound followed by the 2025 squeeze. Revenue recovered to RMB28.663 billion in 2024 and RMB30.994 billion in 2025. Profit climbed to RMB6.241 billion in 2024 before retreating to RMB5.825 billion in 2025. Gross margin rose to 28.1% in 2024, close to the economics investors associated with old Shenzhou, then fell to 26.3% in 2025. The 2025 result therefore contained an important warning: sales were still growing, but incremental sales no longer translated cleanly into incremental profit.

Management gave concrete reasons. Average wages rose. A new Cambodian garment factory had not reached mature efficiency. U.S. tariff changes forced Shenzhou to share part of the tariff burden with customers. RMB appreciation also generated an RMB256 million net FX loss in 2025 against an RMB106 million gain in 2024. These pressures were visible before the much worse H1 2026 result.

The sixth and current stage pits an investment cycle against weaker demand and currency. In six months the workforce rose by more than 9,000 people. The Vietnam asset base increased materially during 2025, Cambodia's asset base continued growing, the second Tay Ninh fabric plant reached 100 tonnes a day, and Indonesia is the next step. At the same time, sportswear orders weakened and major export currencies translated into fewer RMB. Geographical diversification now has to pass a return-on-invested-capital test rather than count as an automatically bullish phrase.

The long financial arc is visible in the four most comparable recent full years:

Sources: company annual-results disclosures.

The historical reputation is visible in the numbers. Even a poor 2022 delivered a mid-teens net margin. Cash conversion held up. Operating cash flow exceeded profit in 2023 and remained close to reported profit across the four-year window. Capital intensity is the weakness: the company must continually invest in textile equipment, environmental systems, automation and new factories to preserve its manufacturing advantage.

FY2025 shows the trade-off. Operating cash flow was RMB5.549 billion. Capital expenditure was RMB3.004 billion, leaving only about RMB2.55 billion of conventional free cash flow, 44% of net profit. Of that capex, 54% went to new factory buildings and prepaid land-use items and 42% to production equipment. With depreciation and amortization around RMB1.49 billion, I estimate maintenance capex at approximately RMB1.3–1.6 billion and growth capex around RMB1.4–1.7 billion. The company does not disclose that split; it is an analytical estimate.

Using a conservative “cash owner earnings” definition of operating cash flow minus approximately RMB1.5 billion maintenance capex, FY2025 owner earnings were about RMB4.05 billion. Today's market capitalization converts to roughly RMB47.1 billion, implying an owner-earnings multiple near 11.6x and an owner-earnings yield around 8.6%. The headline FY2025 P/E is only about 8.1x. That 40%-plus difference in multiples is large enough that I use owner earnings rather than accounting EPS as the anchor for the absolute valuation below.

The balance sheet has strengthened in economic terms despite higher gross borrowings. At the end of 2023, Shenzhou had RMB12.10 billion of bank debt and roughly RMB11.60 billion cash and equivalents, alongside another RMB11.95 billion of other deposit-type financial assets. By the end of 2025, cash and equivalents had increased to RMB15.41 billion, other deposit financial assets were about RMB12.84 billion and borrowings were RMB14.80 billion. At June 2026, the broader deposit-type asset pool still exceeded borrowings by RMB13.33 billion.

Gross debt provides financing and currency matching. Management explicitly states that most sales are settled in USD while procurement is mainly RMB, and that it arranges USD and HKD borrowing to hedge future cash-flow exposure. This explains why a company with enormous deposits also pays finance costs on large borrowings. It does not eliminate FX risk: the RMB506 million H1 2026 remeasurement expense proves that the hedge is partial and accounting outcomes can still be volatile.

Capital allocation has historically been straightforward rather than acquisitive. The 2025 annual results disclosed no significant acquisitions or disposals, while dividends absorbed around 60.9% of annual profit. The proposed FY2025 final dividend of HK$1.20 plus the HK$1.38 interim dividend produced HK$2.58 for the year. When earnings cracked in H1 2026, management reduced the interim dividend to HK$0.88. That behavior is financially prudent: payout follows earning power rather than being defended through leverage.

Four valuation regimes define the capital-market history. The IPO years priced Shenzhou as an emerging Chinese exporter. During the long 2010s, it was re-rated as a quality compounder. The 2020–2021 supply-chain scarcity period pushed that premium to an extreme. Since 2022, the stock has progressively moved back toward manufacturer-type multiples as investors have questioned growth, customer concentration and returns on overseas capacity. A March 2024 BOCI report, for example, showed the stock at HK$65.80 on roughly 20x 2023 earnings; today the TTM P/E is approximately 10.4x.

The historical percentile cannot be stated with false precision because I do not have a complete daily ten-year P/E series in the retrieved source set. Observable anchors put today's valuation roughly in the bottom decile-to-quintile area of the modern quality-compounder period, far below the 2021 extreme. The center of gravity appears to have shifted permanently, at least in part. A 20–30x manufacturing multiple assumed relatively stable high-20s margins and long visibility into customer growth. H1 2026 has broken that stability assumption, at least until proved otherwise.

Business model, moat, industry cycle and horizontal peers

Sportswear orders remain the most important operating variable, as the revenue mix makes clear. H1 2026 sportswear revenue was RMB9.042 billion, 63.8% of sales, down 10.7%. Casual wear increased 4.5% and accounted for 27.9%; lingerie increased 6.1% to 7.0% of sales. The diversification is real, but sportswear is still almost two thirds of the company. A recovery or another leg down in that category can swamp healthy growth elsewhere.

International markets represented 75.6% of H1 sales. Europe fell 8.9%, the U.S. 11.7%, mainland China 5.2%, while Japan increased 2.6%. FY2025 had looked much healthier: Europe grew 20.6%, the U.S. 21.0%, Japan 6.4% and other overseas markets 10.9%, while China fell 8.4%. The abrupt reversal between FY2025 and H1 2026 supports management's description of cautious customer ordering rather than a gradual secular demand decay. It also makes factory planning unusually difficult.

Three layers make up the cost base. Yarn, synthetic-fiber inputs, dyes, trims, energy and transport are substantially variable. Direct labor adjusts only over a longer period because manufacturing plants cannot add or remove thousands of trained workers instantly without damaging delivery capability. Buildings, textile machinery, environmental infrastructure, depreciation and local management are fixed or semi-fixed. Shenzhou therefore has pronounced positive operating leverage when factories fill and equally pronounced negative leverage when orders soften.

Raw materials and consumables charged in H1 2026 increased to RMB6.148 billion from RMB5.883 billion despite lower revenue, while utilities increased to RMB626 million. Management specifically linked higher synthetic-fiber costs to oil. Total employee benefits rose to RMB4.538 billion as well. A factory network designed for rising orders faced higher material, labor and retirement costs while throughput was falling.

Vertical integration is the first genuine moat. Control over knitting, dyeing, finishing and garment construction lets Shenzhou iterate fabrics and garments together, reduce handoffs, maintain color and quality consistency and react more quickly when a brand changes specifications. The financial record is more persuasive than the marketing language: Shenzhou could sustain gross margins in the mid-to-high 20s, sometimes around 30%, while remaining a contract manufacturer. Its 28.1% FY2024 gross margin was exceptionally high for a labor-intensive supplier.

Integration does not confer monopoly pricing. A company with three customers contributing three quarters of revenue cannot unilaterally impose price. The 2025 tariff-sharing disclosure makes that plain: part of the U.S. import-tariff cost ended up inside Shenzhou's gross margin. Customers receive real operational value from Shenzhou but can still force some macroeconomic pain upstream. The relationships are sticky without giving Shenzhou full pricing power.

Scale and accumulated manufacturing know-how form the second moat. Shenzhou can justify specialized textile equipment, automation, fabric development, environmental treatment infrastructure and large multi-country factories that a small cut-and-sew operator cannot. Global brands also need suppliers capable of handling hundreds of millions of units consistently. Scale becomes a liability if the customer base stops growing because expensive assets and a large workforce magnify utilization declines.

Switching friction is the third moat. Shenzhou's customers can move programs, but replacing a large integrated supplier involves fabric qualification, color standards, factory social-compliance audits, production testing, capacity reservation and delivery-risk management. The long persistence of very large customers is the financial evidence. The concentration table also warns against overstating the advantage: the same relationships that create stickiness give a handful of buyers enormous bargaining leverage.

Calculated from company disclosures. The letters are filing labels and should not be assumed to identify the same named brand across years.

For H1 2026 the pattern became still more concentrated:

Source: H1 2026 filing.

Credit terms provide another view of the bargaining dynamics. The company says payment is generally due 30–180 days after delivery. Shenzhou can manage that because of its liquidity, but a large part of the manufacturing cycle is financed before cash collection. Balance-sheet strength is part of the customer proposition as well as a shareholder cushion.

The customer-loss stress test puts concentration directly into the valuation. The calculation below assumes no replacement orders, that lost orders initially carry the current 22.6% gross margin, and an additional 0.5 percentage point gross-margin utilization penalty for every 10% reduction in Customer A orders. That utilization penalty is my scenario assumption, not company guidance.

The exercise is intentionally simple. Customer-specific margins are undisclosed, staffing can eventually be reduced and lost orders may be replaced. It shows the convexity: a 10% cut by the largest account is a 3.4% group-revenue event before utilization effects. A 30% cut is a 10% group-revenue event and could become a roughly one-quarter earnings event if fixed-cost absorption deteriorates at the same time.

Management capability is the fourth moat candidate. Ma Jianrong's organization has spent two decades building an integrated multinational network without relying on transformative acquisitions. The 2025 annual results disclosed no major M&A. Historical cash conversion, dividend discipline and ample liquidity argue for competent capital stewardship.

The present expansion is nevertheless the hardest capital-allocation test in years. China non-current operating assets were RMB5.99 billion at FY2025, Vietnam RMB5.20 billion and Cambodia RMB2.97 billion. By June 2026, China was about RMB5.84 billion, Vietnam RMB5.24 billion and Cambodia RMB2.83 billion. Indonesia will add a fourth meaningful manufacturing geography. Balance-sheet solvency is no longer the risk. Earning an adequate return on those assets is.

Management does not disclose country-level revenue, utilization or return on capital, so an exact Indonesia hurdle cannot be calculated. One analytical boundary is useful: if mature new capacity can earn 15–18% after-tax ROIC at full utilization, it needs roughly 55–70% utilization to clear a 10% cost-of-capital hurdle before allowing for commissioning losses. This is a model, not a management target. Weaker steady-state economics would push the required utilization sharply higher.

The industry is mature and fragmented. Growth depends less on a new end-market penetration curve than on apparel consumption, sports and outdoor mix, brand share shifts, inventory cycles and sourcing reallocation among countries and suppliers. The profit pool is uneven too. Consumer brands capture brand equity and retail economics; manufacturers compete for a smaller share of final retail value. Shenzhou's historical achievement was to capture unusually attractive manufacturing economics inside that structurally buyer-dominated chain.

Several cycles overlap. Brand sell-through drives the consumer cycle, while excess inventories can make procurement fall much faster than end demand. Factory economics move sharply with utilization in the capacity cycle. Raw materials and energy add another layer. Trade policy now adds tariffs that can change sourcing economics faster than factories can move.

Trade policy increasingly follows physical origin. As of the base date, U.S. policy is layered. A July 2025 order set reciprocal rates including 20% for Vietnam and 19% for Cambodia and Indonesia, while China has separately been subject to additional duties under the U.S.-China tariff framework. A February 2026 U.S.-Indonesia deal retained the 19% reciprocal rate but contemplated a zero reciprocal tariff quota for specified textile and apparel volumes linked to U.S. textile inputs.

The picture became more complex again in July 2026. A new U.S. Section 301 action imposed a 10% tariff on goods from Cambodia and Indonesia and directed the creation of textile/apparel tariff-rate quotas tied to imports of U.S. textile inputs; until those quotas are implemented, the 10% Section 301 tariff applies to the specified apparel and textile goods covered by the future quotas. A separate 10% temporary U.S. import surcharge introduced in February 2026 was explicitly limited to 150 days, so it belongs principally to the H1 operating backdrop rather than being assumed permanent in the forecast.

The exact all-in duty on a Shenzhou garment cannot be reduced to “China 10%, Vietnam 20%, Cambodia 19%.” Normal HTS apparel duties, special additional tariffs, product classification, origin and exemptions all matter. The company also does not disclose a weighted tariff rate or the percentage borne by customers. FY2025 merely establishes that Shenzhou shared some incremental tariff cost with customers. That absence of disclosure is itself important: the tariff sensitivity of gross margin cannot be precisely reconstructed from public accounts.

Europe has a different structure. The EU-Vietnam Free Trade Agreement progressively removes EU tariffs on Vietnamese imports through 2027, subject to qualifying rules of origin. Cambodia remains listed under the EU's Everything But Arms regime, which provides duty- and quota-free access for qualifying goods other than arms, again subject to origin requirements. Moving sewing to Southeast Asia is therefore helpful only when the finished product satisfies the applicable product-specific origin test; customer nationality and Shenzhou's Chinese domicile are irrelevant to customs origin.

The most useful horizontal analysis treats peers as different solutions to the same sourcing problem, rather than as a set of ticker multiples.

Crystal International is a broad multi-category apparel platform. Its categories span lifestyle wear, denim, intimate apparel, sweaters, sportswear and outdoor products. That diversification dilutes category concentration relative to Shenzhou. H1 2026 revenue grew 5.4% to US$1.296 billion and net profit increased 10.6% to US$109 million, in direct contrast with Shenzhou's contraction. Crystal's net margin of 8.4% remains much lower than Shenzhou's historical economics, indicating that Shenzhou's integrated fabric model has historically extracted more profit per sales dollar.

Eclat is the strongest "technical fabric plus apparel" comparison. It combines stretch/functional textile capability with garment manufacturing and has historically earned high margins. H1 2026 revenue rose 3.2% to NT$19.71 billion and net profit rose 48.5%; Q2 gross margin reached 30.0%, up from 25.7% a year earlier. Customers choosing Eclat are also buying fabric development and manufacturing know-how, making it the clearest evidence that vertical integration can still command good economics in 2026.

Makalot is more garment-centric and comparatively asset-light. Its economics show the virtue of lower vertical capital intensity: TTM gross margin is around 24.5% and ROIC about 24%, while debt is modest. FY2025 revenue was NT$34.43 billion and net income NT$3.61 billion. It does not offer the same textile integration as Shenzhou, but its higher capital efficiency illustrates the price Shenzhou pays for owning more of the production chain.

The four-company operating snapshot makes the current divergence visible:

Sources use different reporting currencies and reporting periods and are intended as cross-sectional operating indicators, not GAAP-identical comparisons.

Huali Industrial and Yue Yuen are useful adjacent comparisons because they show footwear-manufacturing economics under similar customer and tariff pressures. Huali's FY2025 revenue was RMB24.98 billion with RMB3.21 billion net income, but its H1 2026 attributable profit reportedly fell 43%. Yue Yuen's Q1 gross margin fell 1.2 percentage points to 21.7%, and H1 profitability subsequently weakened sharply. Those results suggest a real regional manufacturing squeeze. They do not excuse Shenzhou's performance because Crystal and Eclat moved in the opposite direction.

The ecological niche is specific. Shenzhou is neither the lowest-cost sewing contractor nor a consumer brand. It is a high-scale integrated technical-apparel supplier that historically offered brand customers unusually reliable product development and production execution. The niche strengthens when brands consolidate orders with fewer sophisticated suppliers. It weakens when those customers deliberately dual-source, demand tariff sharing or migrate volume faster than Shenzhou can resize its fixed-cost base.

The moat verdict is medium-to-strong rather than unambiguously strong. Integration, scale and switching friction remain real. The H1 numbers show that they no longer guarantee high-20s gross margin in an adverse environment. Output per employee and return on new overseas capital are now the proof points, not another automation announcement.

Current fundamentals and the H1 profit bridge

Gross margin is the right place to start reading H1 2026 because the deterioration begins there. Revenue declined by approximately RMB787 million year on year. Gross profit declined by about RMB861 million, meaning the company lost more gross profit than revenue. Cost of sales actually increased slightly despite the smaller top line. This is classic under-absorption behavior in a manufacturing downturn.

Source: H1 2026 interim results.

Operations form the first layer of the profit bridge. Gross profit deterioration of roughly RMB861 million explains more than half the pretax decline. Management identifies RMB appreciation, higher labor expense from wages and capacity expansion, and higher synthetic-fiber costs as principal gross-margin pressures. It quantified the production-labor burden at approximately two percentage points of revenue, equivalent to roughly RMB284 million on H1 sales.

RMB strength also reduced reported sales. H1 2025 revenue was approximately RMB14.97 billion. A 1.8% constant-USD decline would imply H1 2026 sales around RMB14.70 billion before the reporting-currency effect; actual sales were RMB14.18 billion. The implied commercial currency headwind is approximately RMB0.52 billion. This estimate does not mean RMB0.52 billion of profit was lost, because many production costs also move with local currencies. It isolates the revenue translation/conversion effect.

Labor was the other quantifiable operating hit. Total employee benefits reached RMB4.538 billion. Retirement-plan contributions alone increased by RMB55 million, while other benefits increased by about RMB48 million. The new overseas footprint adds more than wage inflation to the cost stack, including statutory contributions, training, commissioning and low initial productivity.

Monetary-item FX remeasurement is the second layer. Net FX expense was RMB506.2 million against an RMB126.4 million gain in H1 2025. The RMB632.6 million swing is almost 10% of H1 2025 pretax profit and about one third of H1 2026 net income. It belongs below gross profit and should not be mixed with the top-line currency conversion effect.

Tax is the third layer. Pretax profit fell approximately RMB1.59 billion. Removing the RMB632.6 million FX remeasurement swing leaves roughly RMB0.96 billion of deterioration from gross profit, other income, operating expenses and financing. Tax expense then fell roughly RMB320 million, partly offsetting the damage. The H1 effective tax rate of around 6.6% was abnormally low relative to 12.5% a year earlier, with deferred tax contributing to that outcome.

One more currency number sits outside profit: exchange differences on translation of foreign operations were an RMB412.6 million OCI loss versus RMB288.1 million a year earlier. That changes comprehensive income and equity translation reserves rather than H1 operating profit. Combining the RMB0.52 billion commercial revenue effect, RMB506 million P&L FX expense and RMB413 million OCI translation number into one “FX loss” would double-count different accounting phenomena.

Tariffs cannot be cleanly quantified in the bridge. The 2025 annual report explicitly says sharing U.S. tariff costs with customers depressed gross margin. H1 2026 was also subject to an unusually volatile tariff environment, including a temporary U.S. import surcharge during much of the half. Shenzhou does not disclose “tariff cost RMB X” or the sharing percentage. Any numerical allocation would be fabricated.

Mix probably contributed, but no reliable amount can be assigned. Sportswear fell from 67.7% of FY2025 sales to 63.8% in H1 2026, while casual and lingerie increased. The company does not disclose gross margin by product, so calling the mix shift accretive or dilutive would require assumptions about customer and product profitability that public filings do not supply.

Working capital is mildly negative. Inventory rose from RMB6.710 billion in December to RMB7.317 billion in June while revenue contracted. On a simple average-balance basis, inventory days rose from roughly 109 in FY2025 to around 116 in H1 2026. Inventory can rise ahead of a second-half production season, and new capacity itself requires more raw material and work in progress, so one half does not prove excess finished goods. Still, weaker orders, higher inventory and more workers raise the evidentiary burden on H2 utilization.

Liquidity remains formidable. Cash and cash equivalents plus current time deposits were approximately RMB26.58 billion at June, and the broader deposit-financial-asset pool was RMB28.10 billion. Bank borrowings were RMB14.76 billion. The broader net-deposit position of RMB13.33 billion is approximately 28% of today's RMB-equivalent equity market value. This materially lowers permanent-loss risk from financing, but cannot prevent permanent loss from overinvesting in low-return capacity.

The external read-through is mixed. Crystal and Eclat reported positive H1 revenue and profit growth while Shenzhou contracted; Huali and Yue Yuen also suffered. Neither extreme interpretation fits. “The whole sector is broken” is false, while “Shenzhou alone has lost its franchise” remains unproven. The evidence supports a company-specific margin and capacity problem inside a difficult broader sourcing environment.

Analyst reactions split accordingly. After the interim result, Jefferies cut its target to HK$36 from HK$47 and maintained Hold, while Nomura reduced its target to HK$63.80 from HK$66.90 and retained Buy. That gap is unusually wide for a mature manufacturer and captures the present argument: one camp capitalizes today's weaker economics; the other assumes normalized utilization and margins return.

Consensus estimates moved down as well. A post-results aggregation showed analysts expecting approximately RMB30.5 billion of FY2026 revenue and RMB3.11 EPS, versus roughly RMB31.8 billion and RMB3.41 before the results. The cuts acknowledge the margin shock but still assume a materially stronger H2 than H1: H1 EPS was only RMB1.27.

The market is trading a margin-normalization probability rather than a simple revenue-growth number. At 10.4x TTM earnings, the price already assumes the old premium is gone. The unresolved question is whether current earnings sit below normalized owner earnings or mark the beginning of a permanently lower return structure.

Four pieces of evidence support the bull case. Shenzhou has previously lifted gross margin rapidly when utilization recovered, including during H2 2023. H1 revenue in USD terms declined only about 1.8%, so reported RMB sales exaggerate the underlying volume/value deterioration. The balance sheet can carry commissioning losses without forcing equity issuance. Casual and lingerie categories still grew while sportswear weakened.

The bear case is equally concrete. The workforce rose 8.5% while sales fell, and top-three customer concentration reached 75.9%. Gross margin had already started falling in 2025 before the H1 FX shock. U.S. trade policy is creating additional costs and rules-of-origin complexity. The next factory is scheduled before current overseas utilization has been publicly demonstrated to have normalized.

The next two reporting periods carry unusual weight. A recovery in revenue without a recovery in gross margin would support the structural bear case. A recovery of gross margin toward 25% with stable customer concentration and improving revenue per employee would provide convincing support for the utilization-trough thesis.

Valuation, margin of safety, risks and tracking

At HK$36.58, Shenzhou's TTM net profit is approximately RMB4.55 billion, comprising H2 2025 profit of roughly RMB2.65 billion plus H1 2026 RMB1.905 billion. TTM EPS is around RMB3.03, or HK$3.53 at the base-date exchange rate. The resulting TTM P/E is about 10.4x, consistent with market-data providers.

The headline multiple is cheap against the old Shenzhou valuation regime, but that does not make it cheap against today's earning power. FY2025 EPS was RMB3.88, which would make the stock only about 8.1x FY2025 earnings. FY2025 also contained considerable growth capex, and H1 2026 proves that accounting profit can outrun cash available to owners when working capital and factory investment rise.

Cash-flow passthrough has been broadly healthy over time. From 2022–2025, cumulative operating cash flow of RMB20.68 billion was 97.6% of cumulative net income of RMB21.19 billion. I estimate the five-year 2021–2025 ratio to be close to 1x; the directly reconciled primary-source four-year window is 0.98x. The business is a genuine cash generator rather than an accrual-heavy manufacturer.

Separating maintenance from growth capex changes the valuation. FY2025 capex of RMB3.00 billion compared with roughly RMB1.49 billion depreciation and amortization. Because 54% of capex funded new buildings/land and another 42% equipment, I use RMB1.5 billion as normalized maintenance capex, with a reasonable range of RMB1.3–1.6 billion. That produces FY2025 cash owner earnings around RMB4.05 billion. The resulting P/owner-earnings multiple is roughly 11.6x and owner-earnings yield 8.6%, versus an 8.1x headline FY2025 P/E. The gap is greater than 30%, so the scenarios below use owner earnings.

Peer valuation settles less than it appears to. Makalot's TTM capital returns are higher and its operating model is less asset-heavy, deserving some premium on capital efficiency. Eclat's current margin performance is superior and deserves a quality premium while that persists. Crystal earns lower margins but is more category-diversified. Huali and Yue Yuen show that manufacturing multiples can remain low when customer and tariff risk dominate the story. Shenzhou should no longer receive a blanket “best manufacturer” premium until its return on new capacity is visible.

I would not use the old 20x-plus center as a historical valuation anchor. The 2021 peak capitalized an unusually favorable combination of high margins, growth, supply-chain scarcity and low perceived disruption risk. A sustainable post-2026 center is more likely around 10–14x normalized owner earnings unless Shenzhou restores high-20s gross margin and high-teens returns on invested capital.

The absolute valuation rests on three sets of conditions. The conservative case assumes H1's margin damage proves sticky. The base case assumes utilization normalizes but the historical premium does not fully return. The optimistic case requires both operating recovery and evidence that international capacity earns attractive returns. All RMB earnings are translated at 1 RMB = HK$1.1664.

These are valuation scenarios within a research framework, not investment advice.

The conservative multiple reflects the prospect that a manufacturer with 75.9% top-three concentration, 22.6% gross margin and continuing capacity commitments can trade at a single-digit owner-earnings multiple if margins fail to recover. The base 10.5x multiple deliberately remains far below the old quality-compounder regime. Even the optimistic 12.5x does not require the 20x-plus valuation investors once paid; most of the upside comes from restoring owner earnings.

Margin recovery from 22.6% toward roughly 25.5% is the base case's most fragile assumption. If only 70% of the assumed recovery arrives, gross margin would be around 24.6–25.0% rather than 25.5%. On my model, owner EPS falls toward roughly RMB3.2, putting base value around HK$39–40 at the same 10.5x multiple. The present price would then offer little capital upside beyond dividends.

Four numbers contain the expectation gap: sportswear revenue growth, gross margin, staff cost as a percentage of sales and top-three customer concentration. Revenue alone can mislead. A zero-growth half with gross margin recovering to 25% could be positive for the stock; 5% sales growth with gross margin stuck at 22–23% would be negative because it would imply the incremental capacity is not recreating old economics.

The margin-of-safety test is less favorable than the headline P/E. Current HK$36.58 is approximately 24% above the HK$29.4 conservative value, leaving zero margin of safety against that scenario. The price does sit near the low edge of the base fair-value zone. Shenzhou is therefore better described as a “good historical company at a price that requires operational stabilization” than as a classic deep-value situation.

The flat-earnings stress test is less alarming. If TTM EPS of roughly RMB3.03 simply remains flat for three years and payout stays near 60%, annual dividends would be approximately HK$2.1 per share at the present exchange rate, a cash yield around 5.8% before tax. I did not retrieve a base-date official 10-year Chinese government-bond close in the final source set, so I will not manufacture an exact comparison. The dividend test itself does not fail. The greater risk is that flat earnings may prove too optimistic if capacity returns decline.

Margin-of-safety sufficiency verdict: none.

The main permanent-loss risk is customer concentration. I assign medium probability and high impact to a material order reduction by one of the top three customers. The observable indicator is each anonymous customer's share and absolute revenue at every interim and annual report. A 20% reduction from Customer A, without replacement orders, removes about 6.7% of group revenue before utilization effects. The transmission path is lower factory loading, lower gross margin, weaker cash conversion and then multiple compression.

The second is structural margin reset, with medium-high probability and high impact. The alert is gross margin remaining below 24% for two consecutive reporting periods even after RMB/USD stops moving materially against Shenzhou. That would imply labor, tariffs, customer pricing or capacity inefficiency is the larger problem. A manufacturer earning 22–23% gross margin deserves a very different multiple from one consistently earning 27–30%.

The third is overseas overbuild. I assign medium probability and high impact. The indicators are headcount growth materially above sales growth, declining revenue per employee, rising overseas non-current assets and continued capex before utilization recovers. If Indonesia opens in late 2027 while Cambodia and Vietnam remain below economic utilization, depreciation and labor expense can turn diversification into a multi-year return-on-capital drag.

The fourth is trade-policy escalation. Probability is high for continued policy changes, while the financial impact ranges from medium to high depending on customer sharing. The July 2026 U.S. Section 301 action and the Indonesia textile quota mechanism show that sourcing advantages can increasingly depend on where fabric inputs originate, not merely where sewing occurs. The key indicator is incremental tariff percentage on the specific apparel HS codes produced in Cambodia, Vietnam and Indonesia and the proportion customers agree to absorb.

The fifth is FX. I assign medium probability and high earnings-volatility impact but lower permanent-loss impact because the balance sheet is liquid and the company actively hedges. The warning threshold is net P&L FX expense above 2% of revenue; H1 2026 was about 3.6%. Persistent RMB strength matters even without a remeasurement loss because it reduces reported USD export revenue relative to RMB costs.

Governance risk is lower but not zero. Founder-family control aligns a large block of equity with long-run value, while also reducing the influence of outside shareholders. The company reports compliance with the applicable Hong Kong corporate-governance code and has disclosed no material contingent liability at FY2025. I found no primary disclosure in the reviewed period indicating an accounting fraud, auditor crisis or material litigation that would override the operating thesis.

Positive catalysts are straightforward: sportswear orders turning positive, gross margin recovering above 25%, a reversal of the RMB appreciation, measurable productivity improvement in Cambodia and Vietnam, tariff-sharing terms becoming less onerous and evidence that Indonesia is linked to incremental customer programs rather than speculative capacity.

Negative catalysts are the mirror image: another gross-margin miss, Customer A/B/C concentration rising above 80%, continued employee growth despite flat sales, an Indonesia capex step-up before current capacity normalizes, further U.S. tariff escalation, or a second half in which operating cash flow materially trails profit.

The expected next full-year results date has not been formally announced by Shenzhou as of the research date. A third-party calendar estimates 30 March 2027, consistent with the company's recent late-March reporting cadence; that date should therefore be treated as estimated until an HKEX board-meeting notice confirms it.

The most useful places to track these variables are the HKEX interim and annual announcements for customer shares, margin, headcount and cash flow; the company's own results materials for capacity; Eclat, Crystal and Makalot disclosures for relative order momentum; and White House/USTR and EU Access2Markets material for origin-specific tariff changes. Share-price movements by themselves are the least useful indicator on the dashboard.

Cross-synthesis, final conclusion, uncertainties and sources

Vertically, Shenzhou has proved one capability beyond serious doubt: it can industrialize sophisticated apparel manufacturing at enormous scale and make that operation generate cash through multiple cycles. The 2008 crisis, pandemic disruptions, 2022–2023 destocking and repeated overseas expansions did not destroy the franchise. The four-year 2022–2025 operating-cash-flow conversion of 0.98x net income is particularly important. The historical economics were not an accounting mirage.

Its success was neither pure luck nor pure industry tailwind. The rise of global sportswear gave Shenzhou a favorable end-market, and China's export manufacturing ecosystem supplied labor, infrastructure and upstream inputs. Management then made the more consequential choice to go deeper into fabric technology and process control, while expanding production into Vietnam and Cambodia before geopolitical diversification became a fashionable investor theme. That combination turned a mature manufacturing industry into a high-return niche for Shenzhou.

Those success factors still exist, but their relative strength has changed. Integrated fabric and garment manufacturing still matters. Scale still matters. A liquid balance sheet still matters. Customer concentration has become more severe. Asian labor is more expensive. Trade policy is explicitly interventionist. Brands have stronger incentives to diversify sourcing. The cost of building a multi-country network must now be earned back through utilization rather than simply applauded as “resilience.”

Horizontally, Shenzhou's real advantage over Crystal is historically superior manufacturing profitability; its advantage over Makalot is deeper vertical integration; its closest conceptual comparison is Eclat, where textile development and garment manufacturing also reinforce each other. H1 2026 exposes the weakness on the other side of that model: integrated factories carry more fixed assets and people when orders turn down. Eclat's 30% Q2 margin and Crystal's positive H1 growth remove the excuse that every apparel OEM was forced into Shenzhou's result.

The market is no longer rewarding past success generously. At roughly 10.4x TTM earnings and about 11–12x my normalized cash-owner-earnings measure, the stock has been de-rated from compounder to manufacturer. That is rational. The shares should not recover to 20x simply because they once traded there. A genuine re-rating requires evidence that new capacity can restore owner earnings per share without requiring proportionately more capital.

I think the market may be misjudging one thing on each side. The bearish misjudgment is treating the entire 4.5-point H1 gross-margin decline as permanent. Approximately RMB0.52 billion of reported sales pressure is associated with currency conversion, production labor rose roughly two points of revenue during a rapid expansion and Shenzhou has previously restored margin quickly when utilization normalized. The bullish misjudgment is assuming that this automatically means a return to 28–30% gross margin. Tariff sharing, more diversified customer sourcing and the enlarged fixed-cost base are structurally different from a simple inventory correction.

For the next year, gross margin matters more than revenue. I would rather see 1% sales growth with 25% gross margin than 8% sales growth with 22.5% gross margin. The first result would show that the new network can earn; the second would show that Shenzhou is buying volume with capacity and customer concessions.

Over three years, the decisive variable is return on the Southeast Asian and Indonesian investment. The company has enough cash to build almost anything it reasonably wants. The investment question is whether those factories produce enough after-tax owner earnings per RMB of capital. A business with strong liquidity can still destroy shareholder value if marginal ROIC falls below the cost of capital for years.

Over five years, the question becomes strategic concentration. If the top three customers still represent around three quarters of sales in 2031, the company will remain economically subordinate to a small group of global buyers despite its scale. If new categories, new brands and the Indonesia footprint reduce that concentration while preserving margin, Shenzhou's moat will have broadened.

Core bull reasons

• H1's 22.6% gross margin resembles prior utilization troughs, and Shenzhou previously lifted H2 2023 gross margin to around 25.8% after a weak first half.
• The underlying H1 revenue decline was only about 1.8% in USD terms, materially better than the reported 5.3% RMB decline.
• Deposit-type financial assets exceeded borrowings by approximately RMB13.33 billion, giving the company enough liquidity to absorb commissioning pain without forced financing.
• Current valuation around 10.4x TTM earnings is far below the old quality-compounder regime and embeds a substantial permanent margin reset.

Core bear reasons

• Top-three customer concentration rose to 75.9% and the largest account alone reached 33.5% of H1 sales.
• Employees increased 8.5% in six months while revenue fell 5.3%, pushing staff cost to 32.0% of revenue.
• Gross margin had already declined 1.8 points in FY2025 because of wages, low Cambodian efficiency and tariff sharing before H1 2026's larger FX shock.
• Indonesia will add capacity from late 2027 before public disclosures have established country-level utilization and ROIC for today's expanded Vietnam/Cambodia network.
• Eclat and Crystal grew profits in H1 2026, so Shenzhou's earnings collapse cannot be attributed solely to an unavoidable industry downturn.

The first three-year pre-mortem is customer loss plus utilization. Suppose in 2027 one of the anonymous top accounts reallocates 30% of its Shenzhou orders, with incremental technical-apparel programs moving to other qualified Asian suppliers such as Eclat or Crystal. Group sales fall about 10% before replacement business, gross margin drops toward 20–21% because new Cambodia/Vietnam capacity is underloaded, and owner EPS falls toward RMB2.1. At 7x owner earnings and the current FX rate, the stock would be around HK$17, more than 50% below today's price. The rival-share shift in this script is speculative; the arithmetic of the customer concentration is not.

The second pre-mortem is a slow squeeze rather than a dramatic customer loss. U.S. tariffs remain costly, RMB stays strong, labor inflation continues and Indonesia begins ramping in late 2027. Revenue remains around RMB30–31 billion but gross margin stalls near 22%. Owner EPS settles around RMB2.4 and the market decides Shenzhou is an ordinary capital-intensive manufacturer worth 8x. That produces a price around HK$22, roughly 40% below today's level, even though the company remains solvent and profitable.

Those scripts clarify the permanent-loss path. It is unlikely to be bankruptcy. The balance sheet is too strong for that to be the central case. The more plausible permanent loss comes from a simultaneous reset in normalized earnings and the multiple investors are willing to pay for those earnings.

The reverse would make Shenzhou a better investment. I would upgrade the fundamental view if gross margin rises above 25% while headcount stabilizes, top-three customer concentration falls below roughly 73%, inventory days normalize, and H2/FY operating cash flow keeps pace with net income. Evidence that Vietnam and Cambodia are producing higher output per employee before Indonesia opens would be especially valuable.

I would overturn the thesis negatively if gross margin remains below 23–24% for two consecutive reporting periods after FX stabilizes; if top-three concentration exceeds 80%; if another 8–10% headcount increase occurs without corresponding revenue growth; or if Indonesia capex accelerates while free cash flow and existing overseas productivity remain weak.

The essential distinction is between a damaged quarter and a damaged economic model. H1 2026 is severe enough that investors should demand proof rather than extrapolate the old Shenzhou. Yet today's share price is also severe enough that a full return to old economics is unnecessary for a respectable outcome. My base valuation assumes only a partial margin recovery to 25.5% and a 10.5x owner-earnings multiple.

At HK$36.58, that leaves the shares close to the low end of a reasonable hold range, but above the value of my conservative case. The balance sheet and historical cash generation keep me from treating the company as structurally impaired. The absence of a conservative-case margin of safety keeps me from calling the current price a buy.

【Company-profile scores】

• Fundamental quality: high
• Growth: medium
• Moat: medium
• Financial soundness: strong
• Management credibility: high
• Valuation attractiveness: medium
• Risk level: medium
• Suitable investor type: value / cyclical / long-term growth

【Investment rating】

• Rating: Hold
• One-line thesis: Margin recovery can create upside from 10x earnings, but 75.9% customer concentration and expanding capacity leave no conservative-case margin of safety.
• Ideal buy price: see dedicated line below
• Acceptable hold price: HK$36–49
• Clearly overvalued price: HK$67.5 and above
• Current-price classification: acceptable hold
• Whether to wait for a better price: yes. I would prefer HK$23.5 or below, or alternatively pay more after gross margin recovers above 25% with stable headcount; waiting risks missing a rapid utilization-driven rebound similar to H2 2023.
• Target holding horizon: 3–5 years
• Expected annualized return: conservative about -2%, base about 10–11%, optimistic about 23–24%, using a three-year horizon, the scenario valuation anchors as exit values and cumulative dividends rather than double-counting additional multiple expansion
• Max-loss risk: approximately 50–55% if a top customer cuts orders materially, gross margin falls toward 20–21% and owner earnings are re-rated to around 7x
• Reassessment-trigger signals: gross margin below 24% for two consecutive periods; top-three customer share above 80%; annualized revenue per employee below RMB250,000 without improvement; broader net-deposit position below RMB5 billion; material Indonesia spending before existing overseas utilization normalizes.

【Ideal Buy Price】20–23.5 HKD Basis: at least a 20% discount to the HK$29.4 value implied by the conservative owner-earnings scenario. This is deliberately much lower than the current market price because “ideal buy” is being defined as a genuine conservative-case margin of safety, not merely a discount to the base case.

【Valuation Range】

• current: 36.58 (close as of 2026-09-01)
• bear (conservative · ideal buy zone): [20.0, 23.5]
• base (fair · acceptable hold zone): [36.0, 49.0]
• bull (optimistic · above the clearly-overvalued line): [67.5, 76.0]

The current price sits just inside the lower end of the base hold band. That is consistent with a Hold rating: the stock is inexpensive against Shenzhou's history and offers meaningful upside if utilization normalizes, but a buyer today is still paying a 24% premium to the conservative fair-value estimate. The attractive-looking 10.4x headline P/E is therefore not enough by itself.

Research uncertainties

The first blind spot is customer identity and continuity. Filings anonymize major customers and do not establish that “Customer A” in one year is necessarily the same legal customer as “Customer A” in another. The concentration ratios are reliable; brand mapping is not.

The second is country-level utilization and productivity. Shenzhou discloses geographic non-current assets, aggregate headcount and selected capacity projects, but not a full China/Vietnam/Cambodia matrix of garments produced, labor hours, utilization, operating profit and ROIC. That prevents a precise calculation of the utilization required for each new plant to clear its cost of capital.

The third is working-capital seasonality. I verified December 2024, December 2025 and June 2026 inventory data, but did not retrieve a reliable primary June 2025 inventory breakdown in the final source set. The roughly 116-day H1 2026 inventory calculation should therefore be treated as a warning indicator rather than proof of abnormal finished-goods accumulation.

The fourth is tariffs. Public policy sources establish origin-specific tariff regimes, while Shenzhou confirms it has shared some tariff costs with customers. Neither source provides a weighted average duty paid on Shenzhou's precise product mix or its customer-sharing percentage. HS-code-level sourcing data would be required to quantify the full exposure.

The fifth is maintenance capex. The RMB1.5 billion figure used for owner earnings is my estimate, anchored to depreciation/amortization and the company's disclosed capex mix. If true maintenance requirements are materially higher, owner earnings and all three valuation cases should be reduced.

Source backbone

The most important primary source is Shenzhou's 25 August 2026 H1 interim-results announcement, which supplies the income statement, product and geographic sales, customer concentration, labor expense, FX, liquidity, capex, employee count and Indonesia timetable.

The FY2025 annual-results announcement supplies the 2024–2025 comparison, customer data, geographic assets, tariff-sharing disclosure, cash flow, debt/hedge structure, workforce, capex and dividend payout.

The FY2023 HKEX results supply comparable 2022–2023 customer concentration, regional revenue, cash flow, debt and foreign-exchange policy.

Current market price and valuation anchors come from MarketWatch, Investing.com and WSJ market data, while the base-date RMB/HKD conversion uses XE/market FX data.

Peer operating evidence comes from Crystal International's investor-relations/results materials, Eclat's financial disclosures, Makalot financial data, Huali's filings/financial data and Yue Yuen's results.

Trade-policy analysis relies principally on White House/U.S. trade actions and the European Commission's Access2Markets material rather than secondary tariff commentary.

Other tickers mentioned

• 02232.HK: Crystal International is the broad apparel-manufacturing peer whose H1 2026 growth contrasts with Shenzhou's contraction.
• 1476.TW: Eclat Textile is the closest listed vertical-integration comparison in functional fabric and apparel.
• 1477.TW: Makalot Industrial provides a more garment-centric, less capital-intensive comparison with high capital returns.
• 300979.SHE: Huali Industrial is an adjacent footwear OEM showing similar 2026 customer, cost and trade pressure.
• 00551.HK: Yue Yuen is the large footwear-manufacturing comparison used to test whether margin pressure is industry-wide.
• 9910.TW: Feng Tay is a relevant Taiwan-listed footwear OEM reference within the wider global branded-sports manufacturing ecosystem.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.