688652.SHG

Jingyi is taking share of an existing, large cake — imported semiconductor fab subsystems — not creating a new market. The ceiling is real and still growing, but it is bounded by a narrow product set and by the fact that Jingyi sells subsystems, not full process tools.

The cake is China's fab build-out plus import substitution. China was the largest semiconductor-equipment market in 2025 at about USD49.3bn, and SEMI projects global 300mm fab-equipment spending to rise 18% to USD133bn in 2026 and a further 14% to USD151bn in 2027. That is a powerful tailwind. But Jingyi sells only a thin slice of that spend — chillers, local scrubbers and wafer sorters — so its served market is a fraction of total wafer-fab equipment. 2025 revenue was just CNY1.426bn against a CNY32.93bn (≈329.3亿) market cap.

It is not creating new demand. Chillers and scrubbers are mandatory, long-established subsystems in every fab; the only "new" dimension is made-in-China qualification displacing imports. By 2022 its domestic scrubber share had reached 15.57% (rank 4) and its chillers were domestically leading, so meaningful share runway remains — inside a defined niche.

The honest ceiling: the served market expands through more Chinese fab capacity, deeper subsystem localization, and a growing aftermarket annuity. That can support a multi-year doubling of revenue, but it is share-of-an-existing-cake growth, not blue-sky category creation. To grow into today's valuation, Jingyi needs broad domestic-fab demand across many customers, not one memory champion. Ceiling: substantial, but bounded.

Yes — doubling revenue within five years is plausible and is roughly the report's neutral case, but it is volume- and share-driven, not price-driven, and it is not guaranteed.

Start from the trajectory: revenue grew from CNY742.3m (2023) to CNY1.026bn (2024) to CNY1.426bn (2025), up 38.95% last year, with 2026 Q1 up 16.13% to CNY392.3m. Even at decelerating rates, that compounding doubles the base inside five years. The report's own 2028 scenarios bracket it: optimistic CNY3.4–3.6bn, neutral CNY2.7–3.0bn, pessimistic CNY2.0–2.3bn — the neutral case roughly doubles 2025 revenue in about three years.

The driver is volume and share, not price. Demand visibility is real: contract liabilities (customer prepayments) more than tripled from CNY406.9m (2023) to CNY1,350.9m (2025), then rose again to CNY1,487.9m by 2026 Q1. The macro pulls the same way — SEMI sees 300mm fab-equipment spending rising 18% to USD133bn in 2026 and 14% to USD151bn in 2027. Growth comes from more Chinese fab capacity plus deeper chiller and scrubber penetration at memory and foundry customers, with CXMT expansion alone a plausible CNY120–300m of incremental revenue by 2028. Pricing is not a lever: company gross margin was flat-to-down at 32.61% and scrubber margins slipped.

The constraint is conversion, not orders. Inventory has swelled to CNY2.51bn (2026 Q1), 55.98% of current assets, with 64.31% shipped but not yet accepted. Doubling the topline is realistic; doubling clean earnings is the harder, less certain task, and the income statement showed it in 2025, when net profit fell 3.26% despite 39% revenue growth.

The credible second growth engine is scrubbers maturing into a true co-equal leg beside chillers — and it exists today, but is still subordinate. The flashier candidates (wafer sorters, hydrogen and harsh-gas treatment) exist only in incubation and are far too small to count yet.

Today Jingyi is essentially a one-engine company. In 2025 chillers generated CNY937.0m (about two-thirds of revenue) at 36.37% gross margin; scrubbers CNY348.1m (about a quarter) at 26.73%; wafer sorters just CNY35.6m at a thin 5.00%; spare parts CNY83.6m and service CNY21.4m. So the realistic "second curve" is the existing scrubber line scaling into a genuine compounding engine — it already sells and is qualified, but its margins slipped and it is only a quarter of revenue, not half.

Genuine optionality exists but is not yet a curve. R&D reached CNY140.0m (9.81% of revenue) with 130 invention patents; lead projects span higher-end temperature control, hydrogen and other harsh-gas treatment platforms, and new wafer-transfer products. One sorter project produced an 800-wafer sample that passed end-customer validation, and a high-flow combustible-gas project formed partial prototypes. These are medium-probability commercializations by 2028, not high — they are customer-qualification businesses, not app releases.

A quieter potential second curve is the aftermarket: as the installed base grows, spare parts and service (a combined CNY105m today, at higher margin) could become an annuity that smooths the lumpy equipment cycle.

Honest verdict: the report itself frames the central risk as Jingyi staying "a single-engine chiller story with expensive optionality bolted on top." The second curve is plausible and already visible in the scrubber line, but its scale-up — and any genuinely new platform beyond it — is still unproven.

The moat is real but narrow: it is qualification lock-in, on-site service density and customization for domestic fabs — not a network effect or a technology monopoly. Over three to five years it most likely widens modestly through installed-base accumulation, but it stays a narrow moat exposed to deep-pocketed foreign incumbents and a crowd of domestic peers.

The advantage is switching cost. Once a chiller or scrubber is validated into a fab's production flow — where temperature stability, exhaust safety and tool uptime feed directly into yield — buyers do not switch casually to save a little money. Jingyi has already cleared that bar at YMTC, SMIC, Hua Hong, Intel Dalian, CanSemi and CXMT; by 2022 its domestic scrubber share was 15.57% (rank 4) and its chillers were domestically leading. Local service teams stationed near customers, fast debugging and a direct-sales model deepen the lock-in, and its scrubbers are engineered to pair with Lam, TEL, Applied Materials, Kokusai, AMEC and NAURA tools.

Why it could widen: geopolitics turns "local support certainty" into a product feature; a larger installed base improves procurement, field data, spare-parts pull-through and service revenue; and net cash funds continued R&D (CNY140.0m, 130 invention patents).

Why it stays narrow: this is not software, so margins do not drift up on their own — 2025 company gross margin was 32.61%, down 0.18pp, with scrubber margins slipping under pricing pressure. Foreign benchmarks (SMC in chillers, Edwards/Atlas Copco in scrubbers) keep the highest-value positions, and domestic peers (AMEC, Kingsemi, Hwatsing, PNC, Shengjian) crowd the same space. A durable but modest moat that widens slowly — it does not compound like a platform.

Limited evidence of a strong self-reinvention gene. Jingyi is a focused, engineering-led subsystem specialist with a broadening R&D base — which gives some adaptive capacity within its domain — but it is a state-controlled, single-domain capital-equipment firm, not a demonstrated serial reinventor, and its treatment of bad news is adequate-and-transparent rather than proactively candid in the Baillie sense.

Adaptive capacity is real but bounded. R&D rose 49% to CNY140.0m (9.81% of revenue) in 2025, with 130 invention patents and lead projects across higher-end temperature control, hydrogen and harsh-gas treatment, and new wafer-transfer equipment; an 800-wafer sorter sample passed end-customer validation. So it iterates and extends — but always inside chillers, scrubbers and handling. There is no track record of pivoting into a genuinely new platform, and the sorter business is still only CNY35.6m.

On mistakes and bad news, disclosure is reasonable. The 2025 net-profit decline of 3.26% (to CNY147.9m) is openly attributed to the R&D surge and lower VAT rebates; the inventory build to CNY2.51bn — 55.98% of current assets, with 64.31% shipped but not yet accepted — is disclosed plainly, as are concentration, receivables and write-down risks. A 2024 governance blemish (an independent director "could not be contacted") was disclosed and resolved in 2025, when the director left and a replacement was elected. Some senior executives also trimmed holdings in 2025.

Honest verdict: a competent, transparent specialist that improves its products iteratively — not a culture visibly built to cannibalize itself or reinvent across domains. State control tends to favor steady execution over bold reinvention. For a long-term growth lens this dimension is a "pass, not a strength."

This is the weakest dimension for a Baillie lens, and it must be stated plainly: Jingyi has no founder-owner-operator and no anchor private owner. It is controlled by the Beijing municipal government — the ultimate controller is Beijing's state-asset regulator (SASAC), acting through a chain of state holding companies including Jingyi Group (京仪集团), with an indirect stake of roughly 28%. Management is a competent professional/state team, not deeply equity-bound owners, and some senior executives reduced their holdings in 2025. Do not mistake state stewardship for owner-binding.

For Baillie's question, the prized setup is a founder with a decades-long horizon and personal skin in the game who will sacrifice today's profit for years five-to-ten. Jingyi does not fit that template. A municipal state-owned enterprise carries mixed objectives — industrial-policy, employment and local-development goals alongside shareholder returns — and professional managers' pay is not tied to a large personal equity stake. The structure is state-linked industrial lineage, STAR-listed, with no founder/anchor owner highlighted.

There are genuine positives. State control aligns the company tightly with the national semiconductor-localization mission, supplies patient capital and audited disclosure, and underwrites a net-cash balance sheet. The willingness to let 2025 R&D jump 49% while net profit fell 3.26% is consistent with spending ahead of profit — though it is partly mandate-driven rather than pure owner conviction.

The negatives a growth investor must weigh: insider selling into a stock up about 252% at roughly 208x earnings is a mild caution flag, and the 2024 independent-director vacancy (now resolved) shows governance is sound but not pristine.

Verdict: long-term-oriented, yes; deeply owner-bound, no. Solid state stewardship — not the founder-owner ideal Baillie hunts.

Jingyi passes both tests, with caveats. On indispensability: if it vanished tomorrow, its qualified customers — YMTC, SMIC, Hua Hong, CXMT, Intel Dalian — would feel real pain, because re-qualifying chillers and scrubbers into running production flows is slow and risky; but foreign incumbents and domestic peers could backfill over time, so it is "hard to replace quickly," not "irreplaceable." On sustainability: clean — its growth helps rather than harms society, and faces no serious domestic regulatory backlash.

Indispensability is moderate-high, not maximal. Subsystems sit close to yield, uptime and safety, and fabs run strict supplier screening, so an installed vendor is genuinely missed. But Jingyi was only rank 4 in domestic scrubbers (15.57% in 2022), with credible alternatives — Edwards (Atlas Copco) and SMC among foreign peers, and AMEC, Kingsemi, Hwatsing, PNC and Shengjian domestically. Customers would scramble, not be stranded.

Sustainability passes the dual test cleanly. Socially, local scrubbers treat toxic process exhaust and chillers improve fab efficiency — its products make chip-making safer and cleaner — and it advances China's supply-chain resilience without any extractive or predatory model. On regulation, domestic policy is a tailwind, and the only material regulatory threat is foreign export controls, which can delay customer build-outs or pinch upstream parts. Tellingly, those controls cut in Jingyi's favor on balance, strengthening the "buy domestic" rationale; and external validation of Chinese memory demand is real, with Micron acknowledging the growth of customers like CXMT and YMTC.

Dual-test verdict: genuinely (if not uniquely) missed, and growth that is societally benign and policy-aligned — it clears Baillie's "good for the world over the long run" bar more cleanly than most growth stocks.

Decent industrial unit economics — but distinctly not software-like, and crucially they did not improve with scale in 2025. Company gross margin was 32.61%, down 0.18pp year on year, even as revenue grew 39%, and net profit actually fell 3.26%. Incremental returns are real but modest and lumpy, gated by a long acceptance cycle, and cash is currently flowing into working capital and R&D rather than into clean profit.

The segment economics show one good engine and two weaker ones. In 2025 chillers earned CNY937.0m at 36.37% gross margin (the quality leg); scrubbers CNY348.1m at 26.73% (margins slipped); wafer sorters just CNY35.6m at a dilutive 5.00%; spare parts CNY83.6m and service CNY21.4m (higher-margin but tiny). The blended margin was flat because, as the company puts it, cost growth broadly matched sales growth — there is no automatic operating leverage here.

Profit quality lagged the topline. Attributable net profit slipped to CNY147.9m (−3.26%) despite +38.95% revenue, because R&D rose 49% to CNY140.0m and VAT rebates fell; non-recurring items (CNY20.75m of subsidies, CNY18.14m of financial-asset gains) still mattered.

Where the cash goes: 2025 operating cash flow turned strongly positive at CNY381.4m — hard to fake in capital equipment — but inventory ballooned to CNY2.39bn (2025) and CNY2.51bn (2026 Q1), 55.98% of current assets, with 64.31% shipped but not yet accepted. So cash is tied up in the acceptance cycle and in R&D; the balance sheet stays net cash (CNY554.3m cash plus CNY942.9m of trading financial assets against CNY50m of short-term debt, falling to zero by 2026 Q1).

Verdict: respectable margins, real but still-unproven operating leverage. It improves only if the mix tilts toward chillers and service and acceptance converts cleanly.

A 10-year 5x from CNY196.01 — to roughly CNY980 and a market cap near CNY165bn (≈1,646亿) — is possible but a low-probability, blue-sky outcome, and it is unattractive on a risk-adjusted basis because today's price already prepays years of flawless execution. The stock trades near 208x trailing earnings and about 23x 2025 sales after a 252% one-year run. The report's own optimistic 2028 case reaches only CNY220–250 (about +12% to +28%), while its neutral case implies CNY100–120 (−39% to −49%) and the pessimistic case CNY45–60 (−69% to −77%). Near-term expected value is negative.

For a 10-year 5x, several conditions must hold together:

• Chinese fab capex keeps compounding for a decade — plausible into 2026–2027, where SEMI sees +18% then +14% to USD151bn, but a full decade is unknowable.
• Jingyi becomes a true multi-engine platform: chillers and scrubbers both compound, sorters and harsh-gas products commercialize, service scales, and gross margin expands from today's flat 32.61%.
• Demand broadens across many customers, not just CXMT — whose expansion (SemiAnalysis projects roughly 500k wafers/month and about 17% of global DRAM supply by 2028, up from ~11% in 2025) must translate into sustained subsystem content for Jingyi.
• Conversion turns clean: contract liabilities (CNY1.49bn) and inventory (CNY2.51bn) become profit and cash, and the market keeps paying a growth multiple.

Effectively revenue must roughly 5–7x with the multiple holding, or 10x+ if the multiple normalizes. Today's price already capitalizes much of this: even the company's own CXMT call adds only about CNY30m of base-case 2028 net profit (≈CNY8–9 per share in a fair-value frame). The path exists; the odds-weighted return from CNY196 does not justify it for a disciplined growth investor.

The framing inverts here, so honesty matters: the market has very much noticed Jingyi — up about 252% in a year to roughly 208x earnings and 23x sales, with daily turnover far above its three-month average and a 52-week range of CNY53.00–197.77. There is no undiscovered-gem mispricing left to harvest on the upside; if anything, the misjudgment now runs the other way, toward over-enthusiasm.

Tested against the three classic reasons the market overlooks a stock:

• Can't-understand: subsystem vendors are unglamorous and their accounting looks messy — revenue runs through shipment then acceptance, so inventory swells before earnings (today 64.31% of inventory is shipped but not yet accepted). Generalists historically under-rank such names. This was the original mispricing — now corrected, and then some.
• Look-down: yes, historically investors ranked chiller and scrubber makers below front-end lithography and etch champions, even though fabs care intensely about temperature stability and exhaust safety. Largely repriced.
• Can't-see-far: the genuine bull case is the years-three-to-ten localization runway plus CXMT optionality (SemiAnalysis sees CXMT heading toward ~17% of global DRAM supply by 2028), which is hard to model. But the market is now over-extrapolating it into "every new Chinese fab benefits Jingyi" — directionally true, not proportionally true.

The narrative inflection point cuts both ways. An upside re-rating needs sustained contract-liability growth combined with cleaner profit conversion: rising deducted profit, chiller margins stabilizing, scrubbers becoming a real second engine, and a credible multi-customer content read. The more likely near-term inflection is downward — if contract liabilities flatten, inventory keeps outrunning sales, or acceptance slips, the stock re-rates from a scarcity multiple toward a normal growth-equipment multiple (the report uses 50–60x), which would gut the price even if the business keeps growing.