688037.SHG

Kingsemi is growing its way into an existing pie, not creating a new market. Its core franchise, front-end coat/develop track, sits inside a profit pool that overseas vendors still overwhelmingly control: Tokyo Electron disclosed in April 2026 that its global coater/developer share exceeds 90%, and Kingsemi's own materials describe domestic localization in this step as still low. The ceiling is therefore defined by two forces, total Chinese fab equipment spending and how much of the import-held profit pool is genuinely contestable. SEMI data frame the scale: worldwide semiconductor-equipment sales reached US$135.1 billion in 2025, up 15%, with China the largest regional market at US$49.3 billion. Kingsemi does not need to displace the global leader everywhere; it only needs to become the domestic default alternative in enough Chinese capacity additions and technology layers, which is why even partial success in a thinly localized bottleneck can carry value far above current revenue.

The one place Kingsemi genuinely enlarges its own addressable pie is process adjacency. Its 2025 materials note that cleaning steps account for more than 30% of all manufacturing steps and rise as nodes shrink and FinFET and GAA structures spread, so a credible front-end chemical cleaner alongside track widens the served market without abandoning the original process-intimacy advantage.

The report lands on a sound industrial logic but a stretched market logic. This is a share-gain-through-import-substitution story bounded by China's capex cycle and by foreign incumbents defending the highest-value layers, not the creation of a brand-new market. At CNY 439.05 and a CNY 88.52 billion market cap, the ceiling is real but already priced as if largely captured.

Doubling revenue within five years is realistic; it is the least demanding part of the thesis. Kingsemi's top line has grown steadily, from CNY 1.717 billion in 2023 to CNY 1.754 billion in 2024 and CNY 1.948 billion in 2025 (up 11.11%), and 2026 Q1 revenue rose 20.08% year on year to CNY 330.6 million on rising new orders and delivery and acceptance progress. The report's own scenarios imply doubling well before year five: 2028 revenue of CNY 3.5 billion (conservative), CNY 4.5 billion (base), or CNY 5.5 billion (optimistic) against the 2025 base of CNY 1.948 billion. The base case alone is roughly 2.3 times by 2028.

Growth is driven mainly by volume and new business, not price. On volume, the front-end coat/develop track continues to ramp as domestic fabs add capacity and Kingsemi moves up the process stack. On new business, the front-end chemical cleaner has advanced from its 2024 launch to customer validation and repeat orders from multiple large domestic customers, and advanced-packaging wet tools such as temporary bonding, debonding, and frame cleaning add further content per fab. Price is a headwind rather than a tailwind: the 2025 annual report disclosed that some products faced market-opening price pressure severe enough to trigger extra inventory impairment.

So the report lands not on doubt about revenue doubling, which it likely will, but on the warning that revenue growth is not the binding constraint. The real question is whether margin and cash conversion follow. A business can double sales and still disappoint if price pressure and working-capital intensity keep owner earnings thin, which is precisely the risk at a CNY 88.52 billion valuation.

The second curve exists today in credible early form: front-end chemical cleaning, with advanced-packaging wet tools as a further adjacency. Kingsemi launched its front-end chemical cleaner in 2024, and by the 2025 annual-report summary it claimed more than 90% process coverage for multiple SPM applications, cleanliness already meeting advanced-node requirements, three core indicators that can benchmark against overseas leaders, successful customer process verification, and, most important, repeat orders from multiple large domestic customers. This is the first sign that Kingsemi may not remain permanently dependent on one flagship process segment.

The opportunity is structurally large. The company's materials note that cleaning steps account for more than 30% of all manufacturing steps and grow in count and importance as nodes shrink and three-dimensional structures such as FinFET and GAA spread. ACM Research frames the prize: US$626 million, or 69.5% of its US$901 million 2025 revenue, came from single-wafer cleaning, showing how large a dedicated cleaning line can become. Beyond cleaning, advanced-packaging tools such as temporary bonding, debonding, and frame cleaning offer a further leg.

But the report is careful. This is still a promising new product, not yet a proven major order line. If the cleaner remains at validation-and-pilot scale rather than converting into stable, material revenue, the whole track-plus-clean platform thesis shrinks back into a single-category story. So the report lands here: the second curve is real and genuinely alters potential earnings power if it scales, but the commercial proof still rests on repeat orders converting into visible revenue, exactly the follow-through the market is paying for at CNY 439.05 and a CNY 88.52 billion market cap but has not yet been shown.

Kingsemi's moat is real but narrow, and its direction over three to five years is genuinely contested. The core advantage is switching cost tied to process validation: front-end coat/develop tools are tightly coupled with a customer's manufacturing process, photoresist material, and lithography-tool integration; verification cycles are long and expensive; and once a tool is validated in-line, it becomes deeply embedded in yield, throughput, defectivity, and line balance, so customers do not casually swap it. The second layer is domestic scarcity in a category still dominated by imports, where Tokyo Electron's coater/developer share exceeds 90% globally, so Kingsemi only needs to be the domestic default alternative to matter. The third is process adjacency: the front-end chemical cleaner extends its presence in the same manufacturing flow.

What is explicitly not yet a moat is scale economics. Kingsemi lacks Tokyo Electron's installed base, SCREEN's cleaning breadth, NAURA's platform scope, and AMEC's R&D mass. AMEC spent about CNY 3.74 billion on R&D in 2025, 30.23% of revenue, against Kingsemi's 13.06% on a far smaller CNY 1.948 billion base. Kingsemi must still spend hard to defend and extend its position, and 2025 showed margins compress when it pushes for market entry.

The report lands on a moat that is strongest in validated domestic track, emerging in cleaning, and weak as a broad platform. It could widen if cleaning wins sustained repeat orders and localization pressure rises, aided by NAURA's ecosystem; it could narrow if price competition intensifies before the installed base is large enough, or if advanced-node validation stalls while the parent absorbs customer mindshare.

Kingsemi's history shows a genuine reinvention gene. Founded in Shenyang in December 2002 as a Sino-foreign joint venture, it spent more than two decades climbing deliberately from smaller-wafer and packaging tools into one of the few domestic platforms shipping volume-production front-end coat/develop tracks, choosing progressively harder, stickier process steps rather than staying in easier niches. It is now attempting its next reinvention, from a single-category track specialist into a broader wet-process and advanced-packaging platform, with the 2024 launch of the front-end chemical cleaner as the leading edge. That pattern, surviving long enough to win expensive validations and then extending into adjacent process steps, is the opposite of a one-product company waiting to be disrupted.

On treating mistakes and bad news, the evidence is mostly reassuring with one caveat. The 2025 annual report candidly disclosed uncomfortable facts: some products faced market-opening price pressure severe enough to require inventory impairment, government subsidies fell, and deducted profit turned negative. More unusually, on June 30, 2026 the company issued its own risk reminder warning that its dynamic P/E was far above the industry, that the stock had risen 78.98% in a month, and that short-term moves reflected overheated sentiment and possible irrational trading, rare candor about its own share price.

The caveat: management's framing that 2025 profit fell mainly because R&D exploded is slightly overstated, since R&D intensity actually fell to 13.06% from 16.92% in 2024. And some information is withheld as commercial secret. The report lands on a company with a demonstrated capacity to reinvent and reasonable honesty about setbacks, tempered by mild narrative spin.

Management has shown clear long-term vision, but the founder-deeply-bound framing no longer fits cleanly after 2025. On vision: Kingsemi spent more than twenty years climbing from a modest Shenyang joint venture into a credible domestic front-end track supplier, deliberately choosing a process step where validation is slow and expensive but stickiness is high, accepting years of delayed payoff. It has sustained R&D above 10% of revenue for years, 13.06% in 2025 and 16.92% in 2024, and absorbed a sharp 2025 profit hit, with attributable net profit down 64.64% to CNY 71.7 million and ROE dropping to 2.62%, rather than starving investment. That is willingness to sacrifice current profit for future capability.

But alignment has fundamentally changed. Before June 2025, Kingsemi had no controlling shareholder or actual controller. Then NAURA acquired 17.87% (9.49% from Advanced Manufacturing at CNY 88.48 per share and 8.41% from Zhongke Tiansheng at CNY 85.71), took four non-independent and one independent board seats, and became controlling shareholder, with Beijing Electronics Holdings as ultimate controller. The original controllers partly cashed out; leadership now reflects the new order, with chairman Dong Boyu and Deng Xiaojun, formerly in North Huachuang's equipment unit, as director and executive-committee chairman.

So interests are now bound less to a founder-owner and more to a state-backed platform parent that booked CNY 2.117 billion of goodwill on the deal. The report lands on demonstrated long-term orientation but flags the governance risk that Kingsemi's stand-alone economics could be optimized for NAURA's broader platform agenda rather than for minority shareholders, something to watch through related-party disclosure and technical-staff retention.

If Kingsemi vanished tomorrow, its front-end track customers would miss it meaningfully, but they would not be stranded. It is one of the few domestic suppliers able to ship production-type coat/develop tools in a category where localization is still low, and its higher-end track has already achieved in-line mass-production work with mainstream lithography tools in complex processes. Because a validated track tool is embedded in yield, throughput, defectivity, and line balance, and verification cycles are long and costly, customers cannot casually replace it. Its disappearance would remove their domestic option and their supply-security hedge against export controls, real value in a fragmenting environment where key-equipment autonomy is a stated national priority.

The important qualifier: the pool most likely to reclaim that share is not a domestic rival but the incumbent foreign vendor that never lost the account, Tokyo Electron, still above 90% global coater/developer share. So customers would revert to imports, losing local content and resilience rather than losing the capability entirely. In cleaning, where Kingsemi is earlier and SCREEN, ACM Research, and other foreign leaders remain strong, it would be missed less.

On whether the growth is sustainable and constructive, yes. Kingsemi's expansion rests on import substitution in a bottleneck step and on genuine engineering validation, not on harming society, exploiting users, or courting regulatory damage; it is policy-aligned rather than policy-endangered. The report lands here: Kingsemi is a scarce and genuinely useful domestic asset whose growth is constructive, but useful-and-would-be-missed is a statement about the business, not a justification for paying CNY 439.05 and a CNY 88.52 billion market cap when foreign incumbents still hold the ultimate fallback.

Right now, Kingsemi's unit economics are getting worse as it scales, not better; operating leverage is running in reverse. Revenue rose to CNY 1.948 billion in 2025 (up 11.11%), yet attributable net profit fell 64.64% to CNY 71.7 million, deducted profit turned negative at about CNY 18 million, and ROE collapsed from 11.24% in 2023 to 8.10% in 2024 and 2.62% in 2025. The causes were rising headcount-driven salary and welfare costs, lower government subsidies, and higher inventory impairment as some products faced market-opening price pressure. The fixed engineering and personnel base is expanding ahead of the gross-profit pool.

Incremental returns are further undermined by erratic cash conversion. Operating cash flow was negative CNY 562.9 million in 2023 despite positive profit, positive CNY 441.8 million in 2024, positive but 76.69% lower at CNY 103.0 million in 2025, then negative CNY 343.5 million in 2026 Q1. On an owner-earnings basis, after a maintenance-capex proxy of roughly CNY 50-60 million, 2025 owner earnings are closer to CNY 40-50 million than the headline CNY 71.7 million, thin against an CNY 88.52 billion market cap.

Where does the cash go? Into R&D (13.06% of revenue in 2025), into hiring, and above all into working capital, inventory and pre-delivery production for an order-driven equipment model.

Whether economics improve at scale is the central unproven question. Q1 2026 gross margin rose 12.40 percentage points year on year, hinting that part of the 2025 damage was timing and mix. But the report's most fragile base-case assumption is precisely that margin normalizes while Kingsemi is still forcing its way into imported-product profit pools. The report lands cautious: better economics at scale are possible but unproven.

For Kingsemi to rise five-fold over ten years from CNY 439.05, an unusually long list of conditions would all have to hold at once. Front-end track localization would have to keep climbing the process stack into advanced logic and memory without further price-led impairment; the front-end chemical cleaner would have to convert validation and pilot repeat orders into a large, stable revenue line rather than staying a promising new product; NAURA's control would have to deliver measurable synergies in procurement leverage, cross-selling, and customer access rather than merely changing the ownership chart; gross margin and owner-earnings margin would have to normalize and expand (the report's optimistic case assumes low-40s gross margin and about 13% owner-earnings margin on CNY 5.5 billion of 2028 revenue); cash conversion would have to mature; and the extreme scarcity premium would have to persist.

The catch is what today's price already implies. At CNY 439.05, a CNY 88.52 billion market cap, roughly 45 times trailing sales and about 1,254 times trailing earnings, the market has already pre-paid an enormous amount, valuing Kingsemi above larger, more proven peers (NAURA about 16 times sales, AMEC about 35 times, TEL about 14 times, SCREEN about 5.2 times, ACM about 8.4 times). The report's own optimistic fair value is only about CNY 320, below today's price; even the bull three-year annualized return is roughly minus 10% to minus 11%.

So a 10-year 5x is not realistic from this entry: it would require the business to exceed even the optimistic operating scenario and the multiple to stay extreme for a decade. The report lands with a margin of safety of zero, an Avoid rating, and an ideal buy zone of CNY 90-110, far below where the price implies success is already achieved.

Here the framing inverts. Kingsemi's problem is not that the market has failed to notice it; it is that the market has noticed too much, too fast. The stock rose 78.98% in a single month to CNY 439.05 by June 30, 2026, and the company itself published a risk reminder warning that its dynamic P/E was far above the industry and that the move reflected overheated sentiment and possible irrational trading. So the can't-see, won't-see, can't-see-far question applies to the downside, not the upside.

What the market has over-noticed is the three-layer bull narrative: domestic substitution in a bottleneck tool category, platform expansion from track into cleaning, and NAURA's control change. What it is under-weighting, the won't-see, is the operating reality: deducted profit is negative, ROE is 2.62%, cash conversion is erratic (2026 Q1 operating cash flow was negative CNY 343.5 million), price pressure already triggered inventory impairment, Tokyo Electron still holds over 90% of the global coat/develop pool, and NAURA's CNY 2.117 billion of goodwill embeds synergy expectations before they are proven. The market is pricing future success as present fact, at a CNY 88.52 billion market cap on CNY 71.7 million of 2025 profit.

The narrative inflection the report watches for is operating proof versus disappointment. The bearish trigger is a quarter where revenue still grows but gross margin slips back toward the mid-30s while cash burn stays severe, revealing that the Q1 2026 12.40-point margin recovery was mix, not a turning point, at which the multiple compresses first and earnings estimates follow. A cooling of the broad domestic semicap trade would hit the highest-multiple names first, and Kingsemi has made itself one of them. The bullish inflection would be cleaner repeat orders converting into material revenue plus cash conversion sustained above 0.8 times.