Precision Tsugami (China) Corporation Limited (1651.HK): Deep Dive
Reporting-period check. The fiscal year ends 31 March, following the Japanese parent. The company reports half-yearly: interim results (April to September) come out in mid-November and annual results in late June. The latest period is therefore the half-year ended 30 September 2026, due around 12 November 2026. It was released on that date. The interim announcement is dated 12 November 2026 and declares an HK$0.90 interim dividend. The company holds no earnings call, so the six results events used below are the results announcements, their management commentary, and the MD&A in the interim and annual reports:
- H1 FY2027 interim results, 12 November 2026
- FY2026 annual results, late June 2026
- H1 FY2026 interim results, 13 November 2025
- FY2025 annual results, 25 June 2025
- H1 FY2025 interim results, 13 November 2024
- FY2024 annual results, 26 June 2024
1. What the Company Does
Precision Tsugami (China) builds small machines that cut metal very accurately. A customer feeds a three-metre steel, brass or titanium bar into one end. Out of the other end come thousands of finished parts: a fuel-injector nozzle, a bone screw, a phone-hinge pin, a connector for a water-cooled AI server. Each is turned, drilled, threaded and cut off without a human touching it, and each matches the last to within a few microns. The company makes the machine, not the part. Its customers are the tens of thousands of Chinese machining shops and component makers who turn bar stock into precision parts for other people's products.
The main product, about 84.5% of sales in the latest half-year, is the CNC precision automatic lathe. In Chinese shop-floor slang it is the zouxinji (走心机), the "sliding-headstock" or Swiss-type lathe. The rest of the range is turret lathes, small machining centres, grinding machines, thread- and form-rolling machines and parts. Everything is made in one industrial district in Pinghu, Zhejiang province, about 100 km southwest of Shanghai (company profile).
Origins. The company is the Chinese arm of Tsugami Corporation of Japan, a machine-tool maker founded in 1937 that has been listed in Tokyo for more than fifty years. The Chinese operation was set up in 2003 (Tsugami America). The decision that shaped today's business came gradually over the following decade: China stopped being Tsugami's cheap factory and became its main market. By 2016, per the Frost & Sullivan figure quoted on the company's own site, it ranked first by revenue among foreign-owned CNC high-precision machine-tool makers in China. Tsugami Japan floated the Chinese company in Hong Kong in September 2017 and kept a majority. Since then the parent has steadily handed over more of the economic weight: it sold down part of its stake to widen the free float (Section 10), and most of Tsugami's group production and profit now sits in China. What the Chinese company used to be was a low-cost assembly site for a Japanese brand. What it is now is a China-domestic machine-tool champion that happens to carry a Japanese name, Japanese engineering heritage and a Japanese controlling shareholder.
The problem it solves. Small turned parts are everywhere: in cars, phones, hydraulic valves, medical implants, connectors and, more recently, AI-server cooling loops and robot joints. The part maker's economics depend on three things: how many parts per hour a machine produces, how few it scraps, and how long it holds tolerance before it needs re-adjustment. A Swiss-type lathe that runs unattended through a night shift and still holds a 5-micron diameter by morning is worth far more than a cheaper machine that drifts. Tsugami's pitch is that its machines hold accuracy for years, are easy for ordinary operators to program, and resell well. The company states this as a design aim: build machines "customers would like to purchase again and again" (company profile).
Why it is hard to copy. Holding micron accuracy while a spindle spins at over 10,000 rpm and the headstock slides back and forth all day is a problem of thermal stability, structural rigidity, guide-bushing design and control software. It is not a matter of assembling bought-in parts. Chinese rivals can copy the layout of a Swiss lathe. What takes decades is a design that keeps tolerance over years of three-shift use at a price a job shop can finance, plus a nationwide service network that keeps it running. Section 3 covers the engineering detail.
A customer example. Take a small contract machining shop in Dongguan that has just won an order for brass quick-disconnect couplings used in liquid-cooled AI server racks (the "AI liquid cooling" demand discussed throughout this report). The coupling is a few centimetres long and has internal threads, O-ring grooves and cross-drilled holes. It must seal under pressure, so the bores have tight tolerances. The shop buys several Tsugami Swiss-type lathes through a local distributor. Each machine takes a 3-metre bar into its bar feeder. The sliding headstock pushes the bar through a guide bushing, so cutting always happens right next to the support point. Main and back spindles let the part be machined on both ends, and live tools drill the cross-holes. A finished coupling drops off every minute or two, around the clock. When the order ends, the same machines are reprogrammed to make phone-hinge pins or hydraulic valve spools. That flexibility is why the machines keep their value: their output is general-purpose small parts, not one customer's product.
2. Business Segments
The company reports one operating segment: making and selling CNC high-precision machine tools. It does disclose revenue by product line, and the lines differ enough in customers and economics to describe separately.
Precision lathes (~84.5% of revenue, H1 FY2027; ~83.5% in FY2026). These are the Swiss-type automatic lathes and precision turret lathes. They are the core of the business: the brand, the installed base and nearly all of the competitive position. Swiss-type machines make long, slender parts from bar stock. Turret lathes handle shorter, chucked parts. Lathes grew fastest in the latest half and gained share of the mix (Zhitong via Investing.com, Nov 2026), because the new demand from 3C (consumer electronics), liquid cooling and robotics is mostly for small turned parts. Main rivals: Citizen Machinery, Star Micronics, Nomura DS, Hanwha, and Chinese makers including Shenzhen Today Standard and Ningjiang.
Precision machining centres (~7.5%). These are small, fast vertical machining centres for milling and drilling prismatic parts: housings, brackets, small 3C structural parts. The line is younger. It grew strongly in FY2025 and FY2026, mainly through 3C and parts work that lathe customers also need (Sina, Nov 2025). It exists separately because milling machines compete in a different market, where Tsugami is a challenger. Its role in the group is cross-selling: a shop that runs Tsugami lathes can buy its milling capacity from the same dealer. Main rivals: Brother Industries, Fanuc (Robodrill), and many Chinese and Taiwanese makers.
Precision grinding machines (~3.9%). These are cylindrical and similar grinders used to finish hardened parts to sub-micron tolerances. The automotive industry accounts for more than half of grinder orders, per the H1 FY2027 results. That makes this line the group's most auto-exposed, and it was the one product line that shrank in the latest half as cars entered a down-cycle (Sohu, Nov 2026). Its role is to give Tsugami a presence in high-precision finishing, a more specialised and less price-driven niche than turning. Main rivals: Japanese and European grinder specialists and Chinese state-group makers.
Other (~4%). Thread- and form-rolling machines, components and service. The rolling machines descend from one of Tsugami Japan's oldest product lines. The service and parts revenue comes from the installed base.
Geographic mix. The business is overwhelmingly domestic. In H1 FY2027, mainland China took about 92% of sales and overseas about 8% (Zhitong via Investing.com). The overseas share has moved around. In FY2024 it was about 16% (Sina, Jun 2024). It fell to about 11% in FY2025 (Sina, Jun 2025), dropped to about 6% in H1 FY2026, then recovered as overseas sales rose by more than half in H1 FY2027. Two channels reach overseas customers. Distributors cover Taiwan. The parent's group companies buy finished machines from Pinghu under a master sales agreement and resell them in Japan, Korea, Europe, the United States and elsewhere (HKEX circular, Mar 2026). Two consequences follow. First, the overseas line partly depends on how the parent decides to supply each market. Tsugami Japan also builds machines in Japan and is ramping a newer plant in India. Second, export volumes depend on trade policy that the Chinese company does not control. The company has not explained the FY2026 dip, so no reason is assumed here.
3. Products and Operations
The Swiss-type lathe and why it matters. In a conventional lathe the bar is clamped and the tool travels along it, so a long, thin part bends away from the tool as the cut moves away from the chuck. In a Swiss-type lathe the headstock grips the bar and slides it back and forth through a fixed guide bushing, so the cut always happens within a few millimetres of support. That lets one machine make long, slender parts (pins, shafts, nozzles, screws, connector bodies) to a few microns on diameter and concentricity. Modern machines add a sub-spindle that takes the part as it is parted off and machines the back end, plus live (rotating) tools for milling flats, drilling cross-holes and threading. A finished, complex part leaves the machine in a single setup. With a bar feeder attached, the machine runs unattended for hours.
Tsugami's lineup runs from basic models for high-volume simple parts to multi-axis machines with B-axis and simultaneous five-axis capability for complex medical and 3C parts. Turret lathes sit alongside for shorter, larger-diameter work. The brand promise is "high precision, high speed, high rigidity" (company profile). In practice that rests on three engineering capabilities:
- Thermal control. Spindles, ball screws and motors heat up over a shift, and steel grows about 11 microns per metre per degree. A machine holding 5 microns must either keep its structure thermally stable or compensate. This is the main difference between a machine that is accurate when new and one that stays accurate.
- Guide-bushing and headstock design. The fit between the bar and the bushing, and the stiffness of the sliding headstock, set the achievable accuracy and surface finish.
- Usability. Most Chinese job-shop operators are not CNC experts. Tsugami's reputation partly rests on machines that are easy to set up and program, and on the dealer network training operators.
What the products are used for (the product-to-industry map a chart can draw on):
| Product | What it makes | Main end industries |
|---|---|---|
| Swiss-type automatic lathes | Long, slender turned parts: pins, shafts, nozzles, connectors, bone screws | Auto (fuel and brake systems, EV drivetrain small parts), 3C (hinge pins, camera and connector parts), liquid-cooling couplings, medical, hydraulics, robot screws |
| Turret lathes | Short, larger-diameter chucked parts | Auto, hydraulics, general machinery |
| Machining centres | Milled and drilled prismatic parts | 3C structural parts, auto, general machinery |
| Grinding machines | Hardened parts finished to sub-micron tolerances | Auto (over half of orders), bearings, hydraulics |
| Rolling machines | Threads and forms rolled, not cut | Fasteners, auto |
The new applications. Two product stories explain the recent shift in the customer base (the shift itself is quantified in Section 4):
- AI liquid cooling. Liquid-cooled server racks circulate coolant through cold plates via quick-disconnect couplings and manifolds. Large AI deployments need very large numbers of these small, leak-tight turned parts. By H1 FY2026, the company estimated it had about 5% of domestic orders for liquid-cooling connector manufacturing equipment (Sina, Nov 2025).
- Humanoid robots. Robot joints use planetary roller screws and harmonic reducers. These contain precision-threaded and ground parts that fit Swiss-type lathes and grinders. The company reported orders from dozens of robot-component customers by H1 FY2026, though this is still small and early (same source).
- 3C. Foldable-phone hinges and new component designs led a 3C capital-spending upturn in 2026 (10jqka business outlook).
Manufacturing. All machines are built in Pinghu in a cluster of plants. The process is mostly precision assembly and testing on top of in-house machining of key structural parts and castings, which are partly sourced. Inputs are castings, spindles, ball screws, linear guides and, most importantly, CNC control systems. CNC panels and related components are bought through Tsugami Japan under a master purchase agreement (HKEX circular, Mar 2026). The company does not make its own controllers, which is a real dependency (Section 8.4).
Capacity events
- Fifth plant (Pinghu). Land of roughly 50,000 m² bought in 2023 for a new plant. Management said in November 2024 that it had started operating and would add 3,000-4,000 machines of annual capacity (Tiger Brokers, Nov 2024). Trade press put total Pinghu capacity at about 22,000 machines a year after the expansion (158jixie).
- Sixth plant (Pinghu). Announced with the FY2026 results. The two new plant buildings involve roughly RMB150m of investment and add about 3,000 machines of assembly capacity a year. Haitong International estimated that at about 20% of existing capacity (QQ News, Jul 2026). The timing moved from "by end-2027" at the June 2026 results to "early 2028" at the November 2026 interim, with groundwork complete (Zhitong via Investing.com).
- Lean management. The November 2026 commentary credits lean production with improving throughput in the existing plants. That is how volume grew faster than floor space in FY2026-27, and it suggests the plants are running close to full.
4. Customers
Who buys. The buyers are many and small. They are mostly private Chinese contract machining shops and in-house machining departments of component makers, concentrated in the Yangtze and Pearl River deltas. The company reaches them three ways (HKEX circular, Mar 2026; company profile):
- Distributors in China and Taiwan. The largest channel. Dealers hold demo stock, arrange financing, install and train, and provide first-line service. Some are long-standing partners that Tsugami ranks by tier (one Hangzhou dealer advertises repeated "diamond" dealer awards: cnblogs).
- Direct sales to major end customers. Large component groups and fast-growing customers in new industries.
- Tsugami Japan group. Buys machines for resale outside China (Section 2).
Two subsidiaries, 14 branches and 12 offices give national coverage for sales and service.
End-market mix and how it shifted. Twelve months ago the auto industry was the anchor customer group. In H1 FY2026, auto-parts customers were about 44% of domestic sales, with 3C at about 8% and pneumatics/hydraulics about 7% (Sina, Nov 2025). By H1 FY2027 the picture had changed. "Emerging industries" (AI liquid cooling, optical modules, humanoid robots) roughly tripled. 3C more than doubled. Together they rose from about 16% to about 38% of domestic sales. Auto declined (Trigram; Zhitong via Investing.com). On the disclosed figures, auto is now roughly a third of domestic sales and no longer the majority driver. That estimate is derived, not disclosed.
How each customer type buys:
- Auto-parts makers (Tier-1/Tier-2 suppliers and their subcontract shops). The plant's engineering or production manager makes the decision, often with the owner's sign-off in private firms. They buy for a programme: a new injector, brake-valve or EV drivetrain part. They judge cycle time, accuracy over time and process capability, which is needed for auto quality certification. Sales cycles are a few months because the part process has to be proven on the machine. Once a part is approved on a given machine and process, the customer is reluctant to change. Re-qualifying a safety-related part costs time and paperwork, so repeat orders for the same programme tend to go to the same model.
- 3C contract manufacturers. Demand comes in waves tied to product cycles (a new foldable-hinge design, a camera module). The buyer is the manufacturing engineering team of a contract manufacturer working for a phone brand. Speed of delivery and ability to ship in volume matter most, because a production line has to be up within weeks of a design freeze. Tsugami's advantage here is that it can deliver hundreds of identical machines quickly from a standardised line, plus its usability.
- Emerging-industry part makers (liquid cooling, robots). Often new entrants, or shops pivoting from older work, buying capacity fast when an order arrives. The decision is usually the owner's. Their criteria are delivery lead time, proven ability on similar parts, and resale value if the boom fades. That last point favours a brand with a liquid used market.
- Hydraulics, medical, general machinery. Steadier, more specification-driven buying. Medical customers in particular judge surface finish and traceability.
Why they choose Tsugami. The specific reasons, in rough order:
- Accuracy that holds over years of multi-shift use (Section 3).
- Strong resale value. A used Tsugami holds its price better than many domestic machines, which lowers the effective cost and makes financing easier.
- Operator familiarity. Thousands of Chinese machinists learned on Tsugami machines, and shops can hire people who already know the control and setup conventions.
- Delivery capacity. Pinghu can ship in volume when a demand wave hits.
- The dealer service network.
Switching costs. These are moderate, not severe. Nothing contractually locks a shop in, and machines are bought one at a time. Practical frictions add up, though: programs, tooling, fixtures and operator training built around one brand, and for auto and medical customers, re-qualification of an approved process. A shop with 50 Tsugami machines will usually buy the 51st from Tsugami unless a rival is much cheaper. That is where the friction runs out. In a downturn, price-sensitive shops will try cheaper Chinese machines for simple parts (Section 5).
Concentration and contracts. Concentration is low. No end customer is material on its own. The largest single counterparty has historically been Tsugami Japan itself, as both main supplier and buyer of exported machines (Zhitong, 2018). The parent's share of sales has shrunk as exports fell to single digits. Almost all sales are spot capital-equipment orders placed through dealers, with no long-term supply contracts. Revenue predictability comes from the order backlog, which lasts a few months, and from the installed base's steady demand for parts and service. It does not come from contracts. In practice, revenue tracks Chinese manufacturing capital spending with a short lag.
5. Competitive Landscape and Moat
Market structure. China is the world's largest machine-tool market, and it is fragmented overall. Swiss-type lathes are a much more concentrated niche. In 2018, CICC estimated Tsugami's share of China's precision automatic-lathe market at 58%, first place, even though its share of China's machine-tool market as a whole was under 1% (CICC via 9fzt, Jun 2018). A recent Chinese financial-media piece quoted about 60% for Tsugami and about 18% for Citizen (Sohu). That source is promotional and its method undisclosed, so it is best read as "still the clear number one", not as a measured figure. The reason share is so concentrated: the Swiss-type lathe is a precision niche where Japanese makers defined the product. Tsugami was the first to build it in volume inside China at Chinese costs, and so it got Japanese-grade machines to Chinese job shops at a price nearer domestic machines than imported ones.
The competitors
| Competitor | Country | Listing | Approx Market Cap | Product Overlap | Relative Strength vs Tsugami China |
|---|---|---|---|---|---|
| Citizen Machinery (unit of Citizen Watch) | Japan | TSE: 7762 (parent) | ~¥570bn for Citizen Watch group (Oct 2026, CompaniesMarketCap) | Swiss-type lathes (Cincom, Miyano) | Strongest technical rival, invented the CNC Swiss-type lathe category (1970s); premium in high-end medical and complex parts; makes in Japan, Thailand and China but priced above Tsugami in China |
| Star Micronics | Japan | Delisted 2026 after a tender offer (Globe and Mail/TipRanks); now Private | — | Swiss-type lathes | Respected in Europe and the US; smaller China presence |
| Nomura DS | Japan | Private | — | Swiss-type lathes | Niche, high-end |
| Hanwha (precision machinery unit) | South Korea | Private (unit of Hanwha group) | — | Swiss-type lathes | Price-competitive foreign brand; strong in Korea |
| Brother Industries | Japan | TSE: 6448 | ~¥1.19tn (Oct 2026, StockAnalysis) | Small machining centres | Category leader in fast compact machining centres for 3C; Tsugami is the challenger here |
| Fanuc (Robodrill) | Japan | TSE: 6954 | ~¥5.36tn (Oct 2026, StockAnalysis) | Small machining centres | Dominant control-system maker; Robodrill competes in compact milling |
| Domestic Swiss-type makers (e.g. Shenzhen Today Standard, Ningjiang, Nanjing Jianke, Xi'an Kitamura, Shandong Xinankai) | China | Listing status not verified | — | Swiss-type and turret lathes | Lower price; improving quality; win simple-part, price-led orders |
The parent, Tsugami Corporation (TSE: 6101, about ¥245-265bn in Sept-Oct 2026 per StockAnalysis), is not a competitor in China. It owns the brand and supplies the designs, and its overseas plants decide where non-Chinese demand is served from (Section 8.4).
Why Tsugami wins or loses against each:
- Against Citizen, Tsugami wins on price and delivery in volume work. Citizen keeps the very top end (complex medical, high-mix aerospace) where a shop pays more for the best machine. In a 3C or liquid-cooling surge, where hundreds of machines must arrive within weeks, Tsugami's local volume production wins.
- Against Star, Nomura and Hanwha, the edge is scale in China. Their China presence is too small to match Tsugami's dealer and service coverage.
- Against domestic makers, Tsugami wins on accuracy retention, resale value and brand among operators. It loses on price for simple parts when shops are squeezed. The H1 FY2027 commentary says price competition among domestic CNC makers continues and that the company cut prices on some models to hold share (Trigram).
- In machining centres, Tsugami is a follower. It wins mainly on cross-selling to existing lathe customers.
Barriers to entry. They are real but not prohibitive. Making a Swiss-type lathe that is accurate when it leaves the factory is within reach of many Chinese makers. Making one that stays accurate for years, in volume, with a service network and a used-machine price that customers trust, takes a long time to build. Fresh capital cannot buy that reputation quickly, because it accumulates one machine-year at a time in customers' shops. The weak spot in the barrier is that domestic makers improve steadily, and China's industrial policy actively favours domestic CNC equipment and control systems (Section 6).
Structural shifts. Four are under way:
- Domestic substitution, pushing up from the simple-part end.
- The changing end-market mix (Section 4), which favours the most capable volume maker because new applications need small parts fast.
- Japanese makers rationalising their own footprints: Star's take-private, and Tsugami's own India expansion.
- Geopolitical sorting, where some Chinese buyers may prefer domestic brands and some export markets may prefer non-Chinese-made machines (Section 8.5).
Moat: Narrow - intangible assets (brand and reputation for retained accuracy), scale / cost advantage (largest Swiss-type lathe maker in China), switching costs (installed-base operator familiarity and process qualification)
The evidence: a share in its niche estimated at well over half for most of a decade, resale values and operator familiarity that reinforce repeat purchase (Section 4), and capacity big enough to win demand surges that smaller rivals cannot serve. It is not Wide. The advantage held its share through the 2018-19 and 2022-23 machine-tool downturns, but in both downturns profits fell with volume. The company still has to cut prices on some models to defend share, which shows that the switching costs are soft at the low end. The thing most likely to erode the moat is domestic Swiss-type makers closing the accuracy-retention gap while state policy steers buyers toward domestic equipment.
6. Industry
What drives demand. Machine-tool demand is derived demand. Shops buy machines when they expect to make more parts, so demand tracks manufacturing capital spending, not manufacturing output. Three forces matter in China now:
- Product cycles in downstream industries. EV powertrains, foldable phones, AI servers and robotics each create new parts that need new machines.
- Replacement. China's installed base built up during the 2010-2011 and 2016-2018 booms is aging. The State Council's large-scale equipment-renewal programme, launched in 2024, gives subsidies and tax incentives to replace old equipment.
- Upgrading. The shift from manual and conventional machines to CNC, and from basic CNC to high-precision CNC, as Chinese parts move up in specification.
Size and trajectory. China's machine-tool industry booked about RMB250bn of revenue in Q1 2026, up about 9% year on year, per China Machine Tool & Tool Builders' Association (CMTBA) data as reported by trade press (Zhizao Online). National Bureau of Statistics data show metal-cutting machine-tool output up about 7% in H1 2026 and about 9% over January-August 2026 (DRCNet mirror). That is a third year of recovery after the 2022-23 contraction. Trade press puts China's Swiss-type lathe market at roughly RMB8.5bn in 2024, growing at a high-single-digit annual rate (Sohu). Treat that figure as indicative only, since the source is a promotional piece.
Position in the supply chain. Tsugami is a capital-goods maker one step above component manufacturing. Upstream it relies on castings, spindles, linear motion parts and, above all, CNC controllers, a market dominated globally by Japanese and German suppliers. Downstream it sells to part makers who supply OEMs in autos, electronics, data-centre hardware and machinery. It earns nothing on the parts its machines make. Its revenue comes when the part maker invests.
Import substitution. This works in two directions. Tsugami is itself a substitution story, replacing imported Japanese Swiss-type lathes with locally built ones at much lower cost. Chinese policy now seeks to substitute domestic brands for foreign-brand machines and, especially, for foreign CNC control systems. "Industrial mother machines" (工业母机) are a named strategic priority, and state funding and procurement preferences favour domestic makers. The effect is felt most in state-owned and policy-linked buyers. Private job shops, Tsugami's core customers, still buy mainly on economics.
Regulation. No product approvals are needed to sell machine tools in China. Two regulatory channels matter. The first is export controls: Japan, the US and the EU restrict exports of high-end multi-axis machine tools and some controllers to China, and Chinese policy restricts some technology flows the other way. The second is industrial policy: subsidies for equipment renewal and preferences for domestic content.
Cyclicality. Machine tools are among the most cyclical capital goods. Shops buy in waves when orders are strong and stop abruptly when they are not, so swings in machine-tool orders are a multiple of swings in manufacturing output. China's market fell sharply in 2012-2015, recovered in 2016-2018, fell in 2019, boomed in 2021, fell again in 2022-23 and has recovered since 2024. Small precision machines usually turn earlier in the cycle than large machines, because consumer-electronics and auto-parts capex responds quickly.
Tailwinds and headwinds
- Tailwinds: the equipment-renewal subsidy cycle; AI data-centre build-out, which creates new demand for cooling and optical-module parts; robotics; and continued upgrading from conventional to precision CNC.
- Headwinds: the auto industry's slowdown after three strong years of EV-led capex (flagged by management, Section 8.1); tariffs and trade friction affecting export-oriented part makers; price competition among domestic machine-tool makers; and the risk that new-industry capex proves bursty.
7. Growth Triggers
Ranked by potential impact on the business, most material first. Every item is a forward-looking statement made by management at a results event.
-
AI liquid cooling as a multi-year demand driver. At the H1 FY2027 results, management called AI liquid cooling "the most certain emerging growth driver" and expected it to contribute through FY2028, alongside optical-module parts (10jqka business outlook; H1 FY2027 interim, 12 Nov 2026). This first appeared in the H1 FY2026 interim (13 Nov 2025), which said liquid cooling and humanoid robots would drive the next fiscal year's growth. It was repeated in the FY2026 annual results (June 2026), so it has now been stated at three consecutive events.
-
Sixth Pinghu plant, about 3,000 additional machines of annual assembly capacity. First announced at the FY2026 annual results (June 2026), with two new plant buildings and roughly RMB150m of investment. Production is now targeted for early 2028, with groundwork complete per the H1 FY2027 interim (12 Nov 2026) (QQ News, Jul 2026; Zhitong via Investing.com). Capital spending for FY2027 is the highest in the company's history.
-
Humanoid-robot components (planetary roller screws, harmonic reducers). Management first flagged orders from dozens of robot-component customers at the H1 FY2026 interim (13 Nov 2025). The theme was repeated at the FY2026 annual results and the H1 FY2027 interim, which names humanoid robots as a source of precision-part demand (Sina, Nov 2025; 10jqka).
-
3C product cycle (foldable-phone hinges and component iterations). The H1 FY2027 interim (12 Nov 2026) expects foldable-phone hinges and component changes to keep driving 3C equipment demand. Analysts cited the same at the FY2026 annual results (June 2026).
-
Automotive structural upgrades replacing volume growth. The H1 FY2027 interim (12 Nov 2026) expects auto demand to shift from volume expansion to structural upgrades, especially EV drivetrains and post-machining of integrated chassis castings. This is the recovery path for the auto down-cycle covered in Section 8.1.
-
Domestic equipment renewal. The H1 FY2027 outlook says domestic equipment renewal shows "solid momentum" as policy expectations turn into actual investment, while tariffs and geopolitical disruption persist (10jqka; 12 Nov 2026).
-
Autonomous-driving controllers and AI hardware. Named in the H1 FY2027 outlook (12 Nov 2026) as new sources of demand for precision housings and parts, alongside robots.
8. Key Risks
8.1 The auto down-cycle is deeper or longer than the new industries can offset. Mechanism: auto parts were the largest single customer group until a year ago (Section 4), and more than half of grinder orders come from autos. Management says the auto industry entered a down-cycle from the second half of FY2026, after three strong years (Sohu, Nov 2026). EV-related capex in China has been heavy, and price wars among carmakers squeeze Tier-1 budgets. If auto capex keeps falling while liquid cooling or 3C normalise, the overlap disappears and volumes fall. Calibration: high probability of a moderate drag. Auto is already declining, and only the magnitude is uncertain.
8.2 The new-industry demand is a burst, not a trend. Mechanism: liquid-cooling and 3C demand comes from job shops buying many machines at once for a specific product wave. These shops are often new entrants buying on credit. If AI-server deployments slow, a cooling-connector design is standardised and consolidated with a few suppliers, or Chinese part makers overbuild capacity, orders can stop abruptly. Overbuilt shops then sell used machines, which competes with new sales. The more the mix shifts to these industries (Section 4), the more the company's demand rests on capex cycles it cannot see far into. Calibration: medium probability, potentially sharp effect. These cycles tend to end suddenly, not gradually. Management's own words give the timeframe it is assuming:
"[AI liquid cooling is] the most certain emerging growth driver... expected to contribute through FY2028." (H1 FY2027 interim outlook, 12 Nov 2026, as summarised by 10jqka; translated)
8.3 Domestic price competition erodes the low end. Mechanism: Chinese Swiss-type makers keep improving, and industrial policy favours them (Section 6). The company has already cut prices on some models to hold share (Trigram). If cuts spread from simple models to the mainstream range, the brand premium that supports resale value and dealer economics weakens. Calibration: high probability of a slow drag. It becomes serious only if a domestic maker matches accuracy retention, which has not happened yet.
8.4 Dependence on, and alignment with, the Japanese parent. Mechanism: Tsugami Japan controls 67% of the shares (Section 10). It owns the brand and the core designs, supplies CNC panels and components, and buys the machines that go overseas, all under connected-party agreements (HKEX circular, Mar 2026). Three exposures follow. Transfer pricing on components and exports is set between related parties. The parent decides which plant serves which export market, and it is investing in India, where its operations have been loss-making during start-up (Sina, Jun 2025). And any disruption in the supply of CNC panels routed through Japan would stop assembly. Calibration: low probability of a sudden break; a standing structural exposure for minority holders.
8.5 Geopolitics between China and Japan, and export controls. Mechanism: this is a Japanese-controlled manufacturer selling about 92% of its output inside China. A diplomatic flare-up between Beijing and Tokyo, a consumer or procurement campaign against Japanese brands, or wider Japanese export controls on CNC controllers or precision machine tools could hit demand, or supply, or both. On the export side, US tariffs on Chinese-made goods reduce the appeal of China-built machines to overseas buyers. Management has so far played down the effect:
"After April, as the Trump administration launched a tariff war... [it] has not, so far, had a large effect on the short- and medium-term demand of Chinese manufacturing." (FY2025 annual results commentary, June 2025, as reported by Sina/Zhitong; translated)
The overseas share then fell to about 6% in H1 FY2026 before recovering (Section 2), so the effect on exports was not nil. Calibration: low-to-medium probability, potentially severe. This is the main tail risk.
8.6 Building capacity into a cycle peak. Mechanism: the sixth plant (Section 3) will come online in early 2028, about four years into an upcycle. Machine-tool companies have a long history of adding capacity at the top. If demand turns in 2027-28, the new floor space adds fixed cost just as volumes fall. Calibration: medium probability. The financial damage is cushioned by a net-cash balance sheet and the modest size of the investment.
8.7 Leadership transition. Mechanism: the long-serving chairman and CEO stepped down in November 2025 (Section 10). A change at the top during a demand shift to new industries raises execution risk in pricing, product priorities and dealer relations. The predecessor stays as a full-time senior adviser, which softens the change but also blurs accountability. Calibration: low probability of serious harm. Results since the change have been the strongest in the company's history.
9. Walk the Talk
Results events used: H1 FY2027 interim (12 Nov 2026), FY2026 annual (late June 2026), H1 FY2026 interim (13 Nov 2025), FY2025 annual (25 Jun 2025), H1 FY2025 interim (13 Nov 2024), FY2024 annual (26 Jun 2024). The company holds no analyst call, so commitments are drawn from results commentary and MD&A. Management issues no numerical guidance, so the record below tracks qualitative commitments only.
| What was guided | When | What happened | Verdict |
|---|---|---|---|
| Keep expanding capacity and launch several new models in the new fiscal year | FY2024 annual, Jun 2024 (Sina) | Fifth Pinghu plant operating by Nov 2024; machining-centre and lathe lineup broadened | Kept |
| New Pinghu plant to add 3,000-4,000 machines of capacity | H1 FY2025, Nov 2024 (Tiger) | Plant in operation; volumes absorbed in FY2025-26; plants now near full, prompting the sixth plant | Kept |
| Respond to price competition through pricing optimisation and more R&D spending, not volume chasing | H1 FY2025, Nov 2024 | Product mix improved toward higher-spec machines, though selective price cuts on some models were disclosed by Nov 2026 | Broadly kept |
| Tariffs not a material threat to demand | FY2025 annual, Jun 2025 | Domestic demand held; overseas share fell sharply in H1 FY2026 before recovering in H1 FY2027 | Mixed |
| AI liquid cooling and humanoid robots to drive incremental growth next fiscal year | H1 FY2026, Nov 2025 (Sina) | Emerging-industry sales roughly tripled by H1 FY2027, reaching about a fifth of domestic sales | Kept, and then some |
| Planned leadership succession with continuity | H1 FY2026, Nov 2025 (Caizhongshe) | New chair/CEO and CFO in place; the next two half-years were records | Kept |
| Sixth plant (two new buildings, ~3,000 machines) operational by end-2027 | FY2026 annual, Jun 2026 | Groundwork done; target now "early 2028" | Pending, slipped by about a quarter |
| AI liquid cooling to contribute through FY2028 | H1 FY2027, Nov 2026 | Not yet due | Pending |
The commitments that tell you something
The emerging-industry call was early and right. In November 2025, auto parts were still nearly half of domestic sales. Liquid cooling was a line item of about 5% of the relevant domestic equipment orders, and robots were a list of pilot customers. Management chose that moment to say:
"AI liquid cooling and humanoid robots are expected to drive incremental growth in the next fiscal year." (H1 FY2026 interim, 13 Nov 2025, as reported by Sina/Zhitong; translated)
Within a year, emerging industries had become a larger customer group than 3C, and together with 3C they made up close to two-fifths of domestic sales (Section 4). The call came while the main customer group (autos) was about to weaken. It was specific, it gave a timeframe, and it was proved right, the best evidence in the record that management reads its order book honestly.
The tariff call was half right. Management's June 2025 assurance (quoted in Section 8.5) was correct for the domestic business, which grew strongly. The export line roughly halved as a share of sales over the next six months. The company never explained that drop in the commentary reviewed, and did not flag it in advance. It is the clearest example of management talking about the part of the business that was doing well.
Plant timing moves, quietly. The sixth plant slipped from "by end-2027" in June 2026 to "early 2028" five months later, with no explanation in the summary commentary. The slip is small. The pattern of restating dates without comment is worth watching as capex rises.
Assessment. Management is conservative in tone and specific in what it does say. It gives no numbers, and when it names a direction (capacity, new industries, succession) it follows through. Its blind spot is the overseas line and anything it does not control, which it tends to under-explain, not mis-promise. On the record available, this is a team that does what it says, phrased more cautiously than optimistically.
10. Ownership, Governance and Shareholder Friendliness
Part 1 - Ownership and control. Tsugami Corporation of Japan (TSE: 6101) is the controlling shareholder. It held about 70.9% (270m shares) in April 2023, when it announced a plan to sell up to 22.47m shares to improve liquidity (Stockstar, Apr 2023). It held about 65.89% at the date of the March 2026 connected-transactions circular (HKEX circular, Mar 2026), and about 67.15% after the company cancelled repurchased shares in May 2026 (Trigram). The rest, about a third, is public float. There is one share class, with no weighted voting rights or pyramid. For a minority shareholder the structure means four things. The parent can pass any ordinary resolution on its own. Only connected-transaction votes, where it must abstain, give minorities real power. The company's technology, brand, key component supply and export channel all sit with that same controlling shareholder (Section 8.4). And as the company buys back and cancels shares, the parent's percentage rises mechanically without it buying anything (Section 11).
Part 2 - Governance. Four points are worth noting:
- Board leadership. The chair is an executive, not independent. Per the 13 November 2025 announcement, Tang Donglei resigned as executive director, chairman and CEO, and Li Zequn resigned as executive director. Both said they would spend more time developing the group's successors and stay on as full-time senior advisers. Dr Wang Xiaokun was appointed chairman and CEO. Tang Donghao (a different person from Tang Donglei) and Li Junying were appointed executive directors. Watanabe Nobuhiro was appointed a non-executive director. Waga Katsuichiro moved from non-executive to executive director and became CFO (Caizhongshe, Nov 2025). Combining chair and CEO, with parent-nominated non-executives on the board, leaves independent directors in the minority on the parent's direction. Simply Wall St flagged board independence as a point to watch (Simply Wall St).
- Related-party transactions. These are material and recurring: master purchase (components, CNC panels, equipment) and master sales (machines for export) agreements with the Tsugami Japan group, renewed through the March 2026 circular and subject to independent shareholder approval (HKEX circular).
- Auditor. Ernst & Young reviewed the latest interim. No auditor change or qualified opinion was found.
- Pay. A share award scheme exists. In March 2026 it was amended to give grantees dividend equivalents (Minichart, Mar 2026). Disclosure reviewed did not show whether awards vest on measurable performance targets.
No insider share pledges were found.
Part 3 - Capital returns
Dividends. Dividends per share by fiscal year:
- FY2024: HK$0.80 (0.40 interim + 0.40 final) (QQ News, Jun 2024)
- FY2025: HK$0.95 (0.45 interim, QQ News, Nov 2024; 0.50 final, QQ News, Jun 2025)
- FY2026: HK$1.50 (0.60 interim, Sina, Nov 2025; 0.90 final)
- FY2027 interim: HK$0.90, payable 12 January 2027, up 50% on the prior interim (10jqka, Nov 2026)
The dividend has grown every year, and nearly doubled over two years, roughly in line with earnings.
Buybacks and dilution.
- Last ~90 days (MoatMap record, since 18 Aug 2026): one repurchase, 100,000 shares on 13 November 2026 at HK$34.02-35.30, about HK$3.45m in total, roughly 0.03% of the shares. It came the day after the interim results.
- Older programmes (searched): shareholders approved a repurchase mandate on 19 August 2024 covering up to about 38.06m shares (10%). The company bought regularly through 2025: for example, 100,000 shares on 14 February (QQ News), 500,000 on 28 February (QQ News), and 105,000 on 18 September 2025 (RCCaijing). It cancelled 7.061m repurchased shares on 19 May 2026. Cumulative cancellations since listing reached about 12.76m shares by the H1 FY2027 report (Zhitong via Investing.com).
Against the mandate, actual repurchases were a fraction of the authorisation. Net effect over about three years: shares outstanding fell from roughly 381m (April 2023) to about 368m (late 2026), a decline of about 3.5%, with no meaningful dilution from the share award scheme.
Verdict: Returns Capital - a debt-free balance sheet pays a dividend that has grown every year, and steady small buybacks have shrunk the share count by about 3.5% in three years.
11. Insider Activities
Over the last 12 months, MoatMap's HKEX record shows one disclosure and no director dealings. On 19 May 2026, the controlling shareholder Tsugami Corporation filed an "Other" movement covering 7,061,000 shares (1.92% of the issued shares) (HKEX DI, 2026-05-19). It was not a market trade. It matches the same-day cancellation of 7.061m shares the company had bought back (QQ/Sina search record of the 2026-05-19 cancellation), which raised the parent's percentage holding without any shares changing hands. No director or officer bought or sold on market in the window. With no open-market buying or selling by anyone, the read is neutral: the only "insider" movement is mechanical, and the only buyer of shares in the period was the company itself (Section 10).
12. Scenarios
Bull case. AI data-centre build-out keeps liquid-cooling couplings and optical-module parts in short supply through FY2028, as management expects. Humanoid-robot actuators move from pilot lines to volume, turning dozens of trial customers into repeat buyers of Swiss-type lathes and grinders. The auto down-cycle turns out shallow because EV-drivetrain and chassis-casting upgrades replace lost volume work, so the old core stabilises just as the new industries mature. The sixth Pinghu plant comes online in early 2028 into an order book that needs it, and Tsugami's ability to ship hundreds of identical machines in weeks wins it the bulk of each new wave. Domestic rivals keep winning simple-part orders but stay out of the high-volume, accuracy-critical work. In two to three years the company is a more diversified, less auto-dependent precision-equipment maker, with a customer base split among cooling, electronics, robotics and autos, a larger and younger installed base generating service revenue, and a dividend that keeps rising.
Base case. Management delivers roughly what it has described. Liquid cooling and 3C keep growing but more slowly as the first big capacity wave is absorbed, while autos decline for another year before levelling off. Robots stay a small but growing line. Price competition at the low end continues, and the company cuts prices selectively to hold share, which it can afford. The sixth plant opens in 2028, slightly later than first promised, and is filled gradually rather than immediately. Exports stay a single-digit share, dependent on the parent's plans. Given the Section 9 record of conservative, specific statements that are kept, the base case is a company that stays the dominant Swiss-type maker in China through a mid-cycle pause, keeps its share, and keeps paying out much of its earnings, with growth that is lumpier than the last two years.
Bear case. Risks 8.1 and 8.2 compound. The auto capex decline deepens just as liquid-cooling demand peaks, because AI-server makers consolidate their coupling suppliers and the shops that bought machines in 2025-26 find themselves with spare capacity. Those shops sell used Tsugami machines, which undercuts new-machine pricing. Risk 8.3 then accelerates: domestic makers, helped by policy, take a bigger share of simple-part demand as shops trade down. Risk 8.6 arrives on schedule, with the sixth plant opening into a falling market. Risk 8.5 sits behind all of this. A Sino-Japanese flare-up or a new export-control round would make buyers wary of a Japanese-controlled brand, and could disrupt CNC component supply through the parent (8.4). The business does not break, but it returns to the pattern of past downturns, with share held and volumes sharply lower.
Sources: TipRanks interim announcement (Nov 2026) · Trigram · Zhitong via Investing.com (Nov 2026) · Sohu (Nov 2026) · 10jqka business outlook · 10jqka interim dividend · Sina H1 FY2026 · Sina FY2025 · Sina FY2024 · Tiger Brokers H1 FY2025 · QQ News sixth plant · Minichart FY2026 results · Company profile · Company IR reports · HKEX circular, Mar 2026 · Caizhongshe leadership change · CICC via 9fzt (2018) · Stockstar stake sale (2023) · Zhizao Online industry data · DRCNet industry data · 158jixie Pinghu capacity · Tsugami America · Market caps: Citizen, Tsugami Corp, Brother, Fanuc, Star Micronics tender