Shanghai Hanbell Precise Machinery Co., Ltd. (002158.SZ)

Deep-Dive Research Report — Industrials / General Equipment (Compressors & Vacuum Pumps)

Report date: 14 June 2026 · Listing: Shenzhen Stock Exchange, A-share · Currency: RMB

Section 1: What the Company Does

Hanbell makes the spinning metal hearts of machines that move heat and remove air. Two product families do almost all the work. The first is the screw and centrifugal compressor - the component that compresses refrigerant inside a commercial chiller, a cold-storage warehouse, a heat pump, or a data-center cooling loop, and the "air end" that compresses air inside an industrial air compressor. The second is the vacuum pump - the machine that sucks a chamber down to near-zero pressure so that a process can happen in a clean, air-free environment, used most heavily in solar-cell crystal pulling and, increasingly, in semiconductor wafer fabrication.

If you have stood inside a shopping mall cooled by a central chiller, eaten frozen food that travelled through a cold chain, or used a chip made on a Chinese fab line, there is a reasonable chance a Hanbell rotor was turning somewhere in the background. The company sells the rotating core, and in many cases the complete machine around it, plus the spare parts and servicing that follow for the machine's 15-to-20-year life.

Founding story and why it matters. Hanbell was founded in 1994 in Taiwan, with mainland Chinese manufacturing beginning in 1996 at Fengjing, in Shanghai's Jinshan district. The founder, Liao Zhen'an, named the company around the idea of building "a national brand of the Chinese people" in a field then dominated by Western and Japanese incumbents - the company's account ties the name to the bells of Hanshan Temple. In August 2008 Hanbell became the first Taiwanese-invested company to list as an A-share on the Shenzhen exchange. That heritage matters for two reasons: it gave Hanbell a Taiwanese precision-machining culture (screw-rotor profiling is a craft of micron-level tolerances) grafted onto a mainland Chinese cost base and a mainland customer network, and it left the company controlled to this day through a Panama holding company (Panama Hermes, ~32.7%) with the founding family as the ultimate controller.

The core value proposition. A screw compressor or a dry vacuum pump is not a commodity casting. Its performance hinges on the geometry of two intermeshing helical rotors machined to tolerances that determine how much gas leaks back past them on every revolution. Get the rotor profile, the bearing design, the sealing and the metallurgy right, and the machine is efficient, quiet and lasts two decades. Get it slightly wrong and it is loud, power-hungry and short-lived. Hanbell's value proposition is that it can build this core in China, to a standard close enough to Bitzer / Atlas Copco / Edwards, but at a price and lead time that a Chinese OEM or fab can live with. In the vacuum-pump business specifically, the pitch is import substitution: for years the dry pumps that evacuate a polysilicon crystal-pulling furnace or a semiconductor etch chamber were almost entirely imported, and Hanbell is the leading domestic alternative.

The product in action. Take a polysilicon ingot maker. To grow a monocrystalline silicon ingot, molten silicon must sit in a furnace under a deep, stable vacuum, with no oxygen contamination, for many hours. The customer needs a dry screw vacuum pump that can pull the chamber down, hold it, tolerate the silicon-monoxide dust that sublimates off the melt without seizing, and run for thousands of hours between services. Hanbell supplies that pump, having engineered the rotor coating and bearing life specifically to survive the dust. By 2021 Hanbell held roughly 70% of the Chinese crystal-pulling vacuum-pump market. That same dominance is why the 2025 solar downturn hit the company so hard, and why the next chapter - selling the harsher-duty version of the same pump into semiconductor fabs and selling more compressors into AI data centers - is the whole investment debate.

Section 2: Business Segments

Hanbell reports by product line rather than by division, but the product lines behave like genuinely different businesses with different customers, cycles and competitors. Approximate FY2025 revenue mix: Compressors (and compressor units) ~72%, Vacuum products ~19%, Parts & after-sales service ~8%, Castings ~1%.

2.1 Compressors and compressor units (~72% of revenue)

This is the cash engine and, lately, the growth story. It spans refrigeration screw compressors and centrifugal compressors (the R series), low-temperature screw compressors for freezing and cold chain (the L series), screw air compressors and air ends (the A series), and a newer generation of permanent-magnet variable-frequency screw, magnetic-levitation centrifugal and air-bearing (gas-suspension) centrifugal compressors.

• What it does: supplies the compression core, or the full packaged unit, into commercial central air-conditioning, industrial process cooling, refrigeration and cold chain, heat pumps, general industrial compressed air, and increasingly data-center cooling. Customers are chiller and air-handling OEMs, refrigeration system integrators, heat-pump makers, and industrial end users, predominantly in China with a growing export tail.
• Core capability: rotor profiling and the magnetic-bearing / variable-frequency drive electronics that make a modern oil-free centrifugal machine efficient. The magnetic-levitation and air-bearing centrifugal compressors are technically the hardest things Hanbell makes - they float the rotor on a magnetic or gas film with no mechanical contact, which removes oil and friction and is exactly what hyperscale data-center cooling wants.
• Why it stands alone: it is the oldest, most diversified, most cyclically-resilient part of the company, tied to construction, industrial activity and now AI infrastructure rather than to the single solar cycle that dominates the vacuum line.
• Competitive position: competes with Bitzer, GEA/Bock, Fusheng, Frascold and RefComp in refrigeration screws; with Atlas Copco, Ingersoll Rand and Fusheng in air ends; and with Danfoss Turbocor and a handful of Chinese players in magnetic-bearing centrifugals. Hanbell's edge is a domestic cost position plus a broad catalogue; it loses on brand prestige at the very high end.
• Role in the group: the margin-and-volume base. In 2025 it grew against the trend and carried the company while vacuum collapsed, and management is positioning its centrifugal/magnetic-levitation line as the AI-data-center growth lever.

2.2 Vacuum products (~19% of revenue)

This is the high-beta, high-margin, high-controversy segment. It contains the P series dry vacuum pumps - screw and Roots (booster) pumps - sold heavily into photovoltaics and, as the next act, into semiconductors.

• What it does: supplies dry vacuum pumps that evacuate process chambers. The dominant historical end market was solar: crystal pulling, diffusion and coating steps in polysilicon and wafer manufacturing. The strategic new end market is semiconductor wafer fabrication, where Hanbell has launched three dry-pump families - PMF (clean processes), iPM/PDM (semi-harsh) and iPH (harsh processes) - that have passed SEMI certification and won qualification at some domestic chipmakers, with batch supply underway.
• Core capability: building a dry pump (no oil in the pumping chamber) that survives dirty, corrosive, dust-laden semiconductor and solar processes for thousands of hours. This took years: rotor coatings, bearing life, thermal management. Hanbell is recognised as the leading domestic screw dry-pump maker, the import-substitution candidate against Edwards, Ebara, Pfeiffer and ULVAC.
• Why it stands alone: completely different cyclicality and customer base from compressors. It rides the solar capex cycle (which turned sharply negative in 2024-2025) and, separately, the semiconductor localisation cycle (which is just starting). The margin profile has historically been the richest in the company.
• Competitive position: ~70% of the Chinese crystal-pulling vacuum-pump market at the 2021 peak, dominant domestically in solar; a challenger, not yet an incumbent, in semiconductor dry pumps where Edwards and Ebara own the installed base.
• Role in the group: the option. When solar was booming it was the profit driver; in 2025 it became the drag; the bull thesis is that semiconductor demand replaces solar and re-rates the whole segment.

2.3 Parts and after-sales service (~8% of revenue)

The recurring, sticky tail. Compressors and vacuum pumps run for 15-20 years and need spares, overhauls and servicing throughout. This segment monetises Hanbell's enormous installed base, carries the highest gross margin in the company, and smooths the cyclicality of new-equipment sales. It exists because the installed base exists; it is structurally counter-cyclical (when customers defer new purchases they service old machines longer).

2.4 Castings (~1% of revenue)

A small, vertically-integrated foundry operation. It mostly exists to secure the supply and quality of the iron castings that go into Hanbell's own compressor and pump housings, with a modest external sale. It is not a strategic profit driver; it is supply-chain control.

Section 3: Products and Business Detail

The full catalogue.

• R series - refrigeration screw and centrifugal compressors. The workhorse for commercial central air-conditioning and process cooling. Includes oil-injected screw compressors, oil-free centrifugal machines, and the advanced magnetic-levitation and air-bearing centrifugal compressors aimed at high-efficiency and data-center liquid-cooling duty.
• L series - low-temperature screw compressors. Built for freezing, cold storage and cold-chain logistics, where evaporating temperatures are deep below zero and ordinary compressors struggle.
• A series - screw air compressors and air ends. General industrial compressed air. Hanbell sells both complete air compressors and the bare "air end" (the screw block) to other compressor assemblers.
• P series - vacuum pumps. Dry screw pumps and Roots boosters. The solar-duty versions evacuate crystal-pulling and coating chambers. The semiconductor-duty versions are the PMF / iPM(PDM) / iPH families, segmented by how corrosive and dirty the target process is - PMF for clean steps, iPM for semi-harsh, iPH for harsh etch/deposition chemistries.
• Castings. In-house iron castings for housings, plus modest external sales.

Technical specifications that matter. A Hanbell screw dry vacuum pump reaches an ultimate pressure below ~0.0075-0.01 Torr at a relatively low 2,800 rpm rotation speed - low speed matters because it means less wear and longer service life in dusty duty. The semiconductor pump families carry SEMI certification, the industry credential a fab requires before it will let a pump touch a process tool. For compressors, the differentiating spec is the magnetic-bearing / air-bearing centrifugal design that removes oil from the refrigerant loop and lifts efficiency - the feature data-center operators care about because cooling is a large share of an AI facility's power bill.

Manufacturing. Headquarters and primary manufacturing sit at Fengjing, Jinshan, Shanghai, with branch operations across China (Nanjing, Guangzhou, Chongqing, Jinan) and subsidiary operations reaching Hong Kong, Taiwan, Vietnam, Korea and Indonesia. The company runs its own foundry (the castings segment) to control the quality of the iron that becomes compressor and pump bodies - a real constraint in this business, because casting porosity ruins a precision rotor housing. The manufacturing chain is therefore unusually vertically integrated: cast the housing, profile the rotor, assemble the core, package the unit, service it for two decades.

Geographies and exports. Products reach roughly 50 countries. The home market is China across HVAC, cold chain, solar and semiconductor; the export tail runs through Southeast Asia (Vietnam and Indonesia operations), Korea, Taiwan and broader global distribution for compressors and air ends.

Milestones that changed the business. The 1996 mainland manufacturing start; the 2008 Shenzhen listing (first Taiwanese-invested A-share); the 2010 National High-Tech Enterprise recognition; the build-out of the solar vacuum-pump franchise to ~70% domestic crystal-pulling share at the 2021 peak; and most recently the SEMI certification and first batch supply of semiconductor dry pumps, plus the launch of magnetic-levitation centrifugal compressors aimed at AI data-center cooling - the two pivots the company is now betting on to replace the solar downturn.

Section 4: Customers

Who buys. Three broad customer types. First, equipment OEMs and system integrators - chiller makers, air-handling and refrigeration system builders, heat-pump manufacturers, and other air-compressor assemblers who buy Hanbell air ends - mostly Chinese, with exports. Second, process manufacturers - solar polysilicon/wafer makers and, increasingly, semiconductor fabs and semiconductor equipment makers (Hanbell sells the pump into both the fab and the tool builder). Third, industrial end users and the after-market buying spares and service.

Who decides, and on what. For an OEM buying refrigeration screws or air ends, the decision sits with engineering and procurement, and the criteria are efficiency (coefficient of performance), reliability over a 15-20 year life, price, and lead time. The sales cycle is a qualification process: the compressor must be designed into the OEM's product, tested, and certified, after which it tends to stay designed-in for years. For a solar or semiconductor fab buying vacuum pumps, the buyer is process engineering plus equipment procurement, and the gating criterion is qualification - the pump must survive that specific process chemistry without contaminating the chamber or failing early. In semiconductors that means SEMI certification first, then on-tool qualification at the fab, a slow and exacting cycle measured in quarters.

Why they choose Hanbell. For domestic Chinese customers the answer is a combination of close-enough performance to imported incumbents, a materially better price, far shorter lead times, local service, and - increasingly - a policy and supply-chain preference for a domestic supplier. In solar specifically, Hanbell built its dominance on a cost-performance advantage enabled by its permanent-magnet motor technology and hardened core components that reach service lives competitive with international pumps.

Switching costs. High once designed-in. A compressor or pump that has been engineered into an OEM product or qualified on a fab tool cannot be swapped casually - requalification costs time and risk. The installed base then locks in the spare-parts and service revenue for the life of the machine. This is the quiet moat: not that any single sale is sticky, but that two decades of after-market follows each one.

Concentration. The bigger concentration risk is not a single customer but a single end market - solar. The vacuum line's reliance on the photovoltaic capex cycle is what turned a profit driver into a drag in 2025. The compressor business is more diversified across HVAC, cold chain, industrial and now data center.

Contract structure. A mix of project-based equipment orders (lumpy, tied to customer capex cycles), design-in OEM supply relationships (more durable), and recurring after-market parts and service (the predictable layer). Revenue predictability is therefore moderate: the new-equipment lines swing with their respective capex cycles, while parts and service provide a stable floor.

Section 5: Competitive Landscape

Hanbell competes in two largely separate arenas, against very different opponents.

In vacuum pumps, the incumbents are the global vacuum majors: Edwards and Leybold (both now inside Atlas Copco), Pfeiffer Vacuum (inside Germany's Busch group), Ebara of Japan, and ULVAC of Japan. These firms own the semiconductor dry-pump installed base worldwide. Hanbell's win condition is import substitution: it is the leading Chinese domestic dry-pump maker, it already dominates Chinese solar crystal-pulling vacuum, and it is attacking the semiconductor segment from below on price, lead time and local support. It loses where a fab demands a proven, globally-supported pump with a decade of on-tool reliability data behind it - that is the incumbents' fortress, and Hanbell is still early in qualification there.

In compressors, the refrigeration-screw competitors are Bitzer (Germany, private), GEA/Bock, Fusheng (Taiwan), Frascold and RefComp; the air-end competitors are Atlas Copco, Ingersoll Rand and again Fusheng; and the magnetic-bearing centrifugal opponents include Danfoss Turbocor and a cluster of Chinese players. Hanbell wins on domestic cost, catalogue breadth and increasingly on its magnetic-levitation centrifugal technology for data-center cooling; it loses at the prestige high end where Bitzer and Danfoss command a brand premium and the deepest efficiency track record.

Barriers to entry are real but not impassable. Screw-rotor profiling, magnetic-bearing control electronics, and dry-pump survival engineering each took Hanbell years to build, and SEMI certification plus fab qualification is a multi-quarter gate that protects whoever clears it. But these are engineering barriers, not regulatory monopolies; a well-capitalised entrant with the right talent can scale, and within China several domestic players are chasing the same import-substitution prize. Margins are healthy but not fortress-like, and the solar episode showed how fast a dominant share in one end market can de-rate when the cycle turns.

Structural shift. The dominant force is Chinese localisation: the national push to replace imported semiconductor and vacuum equipment is the tailwind Hanbell is riding in vacuum, while AI data-center buildout is the new pull in compressors. Both are positive structural shifts for a domestic challenger; both also invite domestic competition.

(Market caps are approximate peer-size references only, as of June 2026, and move daily. "—" = privately held.)

Section 6: Industry

What drives demand. Hanbell sits at the intersection of several distinct demand cycles. Compressors are driven by commercial construction and HVAC, cold-chain logistics expansion, the heat-pump transition in heating, industrial activity, and - the newest and most talked-about driver - AI data-center cooling, where rising rack power densities force operators toward high-efficiency and liquid-cooling systems that need efficient compressors. Vacuum pumps are driven by two separate forces: the photovoltaic capex cycle (solar polysilicon and wafer capacity additions) and the semiconductor capex-plus-localisation cycle (new Chinese fab capacity and the policy drive to replace imported vacuum equipment).

Industry size and trajectory. Third-party market research puts the global dry vacuum-pump market in the low-to-mid hundreds of billions of RMB with a low-single-digit to mid-single-digit CAGR (one cited report frames dry-scroll vacuum pumps growing ~5-6% annually through 2033). These top-down figures vary widely by definition and should be read as directional. The more relevant sub-trend for Hanbell is the semiconductor dry-pump localisation theme: vacuum pumps are core consumables in PVD, CVD, etch and ion-implant process systems, and the share currently imported is high, which is precisely the gap Chinese substitution aims to close - brokers covering the stock (Dongwu, Everbright, Huatai) frame semiconductor vacuum and data-center cooling as the two structural growth vectors.

Where Hanbell sits in the supply chain. It is a component-and-equipment supplier one layer below the system level - it sells the compressor into the chiller maker and the pump into the fab or tool builder. In solar vacuum it sits at the top of the domestic supply chain (the dominant local pump); in semiconductor vacuum it is an emerging domestic challenger to imported incumbents.

Import substitution. This is the central industry dynamic for the vacuum line. Chinese fabs and equipment makers historically bought Edwards/Ebara/Pfeiffer pumps; the national localisation drive is opening the door for a qualified domestic alternative, and Hanbell is the most advanced domestic dry-pump candidate (SEMI-certified, in batch supply at some domestic chipmakers).

Regulation and certification. The binding gates are technical certifications rather than government licences: SEMI certification and on-tool fab qualification for semiconductor pumps, plus efficiency standards and refrigerant-transition regulations (the phase-down of high-GWP refrigerants) that shape compressor demand toward more efficient, oil-free designs.

Cyclicality. High and multi-layered. The vacuum business is sharply cyclical, tied to the boom-bust of solar capex (vividly demonstrated by the 2024-2025 solar downturn that roughly halved vacuum revenue). The compressor business is moderately cyclical, tied to construction and industrial capex but cushioned by diversification and by the counter-cyclical after-market. The blended result is a company whose earnings can swing hard when one end market rolls over.

Tailwinds and headwinds at the industry level. Tailwinds: semiconductor localisation, AI-driven data-center cooling demand, the heat-pump and cold-chain build-out, and refrigerant-efficiency regulation favouring advanced compressors. Headwinds: the solar overcapacity and capex contraction that is depressing vacuum demand, intensifying domestic competition in every localisation niche, and currency volatility on the export book.

Section 7: Growth Triggers

Drawn from the five most recent management-on-record disclosures (investor relations activity records and the annual online results briefing filed to the Shenzhen exchange).

• Semiconductor vacuum-pump revenue expected to roughly double in 2026. Investors at the FY2025 results briefing specifically referenced management's prior statement that 2026 semiconductor vacuum-pump revenue could double. (2025 annual online results briefing, 12 May 2026)

• Three semiconductor dry-pump families - PMF, iPM/PDM, iPH - are SEMI-certified, qualified at some domestic chipmakers, and in batch supply. This moves the semiconductor vacuum line from sampling to volume. (2025 annual online results briefing, 12 May 2026; repeated from earlier records)

  "PMF、PDM、iPH三大系列干式真空泵通过SEMI认证，已通过部分国内芯片制造商认可并批量供货。" (The three dry-pump series passed SEMI certification, won recognition from some domestic chipmakers, and are in batch supply.)

• Data-center cooling compressors (permanent-magnet variable-frequency screw, magnetic-levitation centrifugal, air-bearing centrifugal) growing into AI compute demand. Management positions a full product line for AI-driven high-efficiency data-center cooling. (Jan 2026 IR record, 7-15 Jan 2026; repeated at 12 May 2026 briefing)

  "针对数据中心领域，公司有永磁变频的螺杆式、磁悬浮离心式、气悬浮离心式制冷压缩机产品布局。" (For data centers, the company has a product line of PM variable-frequency screw, magnetic-levitation centrifugal and air-bearing centrifugal cooling compressors.)

• Data-center screw and magnetic-levitation centrifugal compressors already grew in H1 2025, indicating the trigger is in execution, not just plan. (Sep 2025 IR record, 23 Sep 2025; reiterated Jan 2026)

• Semiconductor localisation framed as a new growth engine as domestic chip capacity scales, with management explicitly calling current semiconductor contribution small but the runway large. (Jan 2026 IR record, 7-15 Jan 2026; repeated 12 May 2026)

• Compressor business "blossoming at multiple points" across cold chain, heat pumps, equipment-replacement demand and data center - the diversification that carried the company through the 2025 solar trough. (2025 annual online results briefing, 12 May 2026)

• Q1 2026 return to profit growth (net profit up modestly year on year) cited by management as evidence the worst of the solar-driven decline is passing. (post-Q1 IR record, 24 Apr - 29 May 2026)

Section 8: Key Risks

• Solar concentration in the vacuum line - already materialised, may not have fully bottomed. The vacuum segment was built on ~70% domestic share of solar crystal-pulling pumps. When the photovoltaic capex cycle turned, vacuum revenue roughly halved in 2025 and dragged group profit down sharply, with the segment's gross margin also falling. The mechanism is direct: solar over-capacity → polysilicon/wafer makers stop adding capacity → no new pumps bought. Management has acknowledged this as the central cause of the 2025 decline. The risk is that the solar trough lasts longer than the semiconductor ramp takes to fill the hole - a high-probability moderate-to-significant drag rather than a tail risk.

• Semiconductor ramp is a promise, not yet a proven business. Management itself repeatedly says semiconductor vacuum is currently a small contributor. The "doubling in 2026" is off a low base, and fab qualification cycles are slow, exacting and reversible. If domestic fab capex slows, or if Edwards/Ebara defend their installed base aggressively, the re-rating thesis stalls.

  Management's own framing across multiple records: semiconductor vacuum use is "still small, with limited contribution to overall results" - the upside is real but unbanked.

• Data-center cooling is a crowded, contested opportunity. Magnetic-levitation centrifugal cooling is exactly where Danfoss Turbocor, global majors and several well-funded Chinese players are converging. Being early with a product line does not guarantee share; this is a moderate-probability competitive-erosion risk on the most-hyped part of the story.

• Earnings cyclicality and operating leverage cut both ways. 2025 showed how fast profit can fall (net profit down ~46%, far more than revenue, as fixed costs and a richer-margin product mix unwound). The same leverage that punished 2025 helps in a recovery, but it makes the stock's earnings inherently volatile.

• Currency exposure on the export book. 2025 results were further dented by foreign-exchange losses (notably USD volatility), which inflated financial expense. With ~50-country exports, FX swings are a recurring, if second-order, drag - specific enough to flag because it was a named contributor to the 2025 profit miss, not a generic boilerplate risk.

• Controlled ownership and recurring insider selling. The company is controlled through a Panama holding vehicle with the founding family as ultimate controller; the float is concentrated (top-ten holders ~67% of float). Periodic director/executive reduction plans (see Section 11) are a governance-and-overhang risk rather than an operational one, but they shape the supply of stock.

Section 9: Walk the Talk

The five records used (China A-shares hold no quarterly earnings calls; these are the official IR activity records / annual results briefing):

1. 23 Sep 2025 IR record (around H1/Q3 2025)
2. 17-31 Dec 2025 IR record
3. 7-15 Jan 2026 IR record
4. 12 May 2026 - FY2025 annual online results briefing (most recent; within 90 days of the report date)
5. 24 Apr - 29 May 2026 IR record (post-Q1 2026)

The single clearest test of management's word is the semiconductor vacuum-pump claim, because it was made repeatedly and is datable. Through late 2025 and into early 2026 management consistently told investors that semiconductor vacuum was small today but on a steep ramp, and pointed to a specific expectation that 2026 semiconductor vacuum-pump revenue could roughly double.

Across the Sep 2025 and Jan 2026 records: semiconductor vacuum use is "still small, contribution limited," but "with the acceleration of domestic semiconductor localisation and market expansion, the outlook is very broad and could become a new growth driver."

That message was internally consistent across four straight records - management did not oscillate, did not quietly drop it, and at the 12 May 2026 results briefing fielded a direct investor question that referenced its own earlier doubling guidance and reaffirmed the SEMI-certified, batch-supply status. Consistency of message is a point in management's favour.

On the compressor / data-center claim, management said in the Sep 2025 and Jan 2026 records that data-center screw and magnetic-levitation centrifugal compressors had already grown in H1 2025, and laid out a full product line for AI cooling. This is a promise partly evidenced rather than purely forward-looking - they pointed to realised growth, not just a roadmap, which is the more credible form of guidance.

Where management has to be judged more harshly is on the downside. The 2025 collapse in vacuum revenue and the ~46% profit fall were larger than the prior narrative implied, and the solar downturn's depth was not pre-flagged with any precision in earlier communications. To their credit, management did not hide the cause afterwards - they attributed the decline plainly to solar weakness, the vacuum margin drop, and FX losses, rather than blaming the macro vaguely. The Q1 2026 return to modest profit growth, which management then pointed to as evidence of stabilisation, lends some support that they read the trough roughly right.

Assessment. This is management that communicates consistently and owns bad news after the fact, rather than one that overpromises and then spins. The credible pattern is consistent forward narrative on the two pivots (semi vacuum, data-center cooling), plus honest post-hoc explanation of the solar miss. The open question - unanswerable until 2026-2027 results land - is whether the "semiconductor revenue doubling" guidance proves accurate or optimistic. So far the qualitative claims (SEMI certification, batch supply, data-center growth) have been backed by named, datable evidence rather than left as slogans.

Section 10: Shareholder Friendliness Index

Dividends. Hanbell pays a cash dividend every year and the trend tracks earnings. For FY2023 it paid RMB 5.8 per 10 shares (RMB 0.58/share, paid mid-2024); for FY2024 it held the payout flat at RMB 5.8 per 10 shares (RMB 0.58/share, paid mid-2025); and for FY2025 it cut to RMB 4.5 per 10 shares (RMB 0.45/share, paid June 2026), reflecting the ~46% drop in net profit that year. The cut is consistent with a roughly stable payout ratio (FY2025 dividend is around a third of earnings per share) rather than a signal of distress - management lowered the absolute dividend in line with lower profit while keeping the company comfortably profitable and cash-generative (operating cash flow actually rose sharply in 2025).

Buybacks and dilution. No MoatMap database block was injected for this prompt, so the last-90-day window is not separately sourced; on a full three-year web search I found no evidence of a large, executed open-market share-buyback-and-cancellation programme over FY2023-FY2025. Capital return has run through the cash dividend, not repurchases. Share count has been broadly stable, with only modest movement from historical equity-incentive activity rather than meaningful new dilution; I did not find a major buyback authorisation executed in this window. (This should be read as "none located after searching," not a guarantee of zero.)

Verdict: Returns Capital (moderate) - a consistent annual dividend payer that flexes the payout with earnings, with capital return delivered through dividends rather than buybacks.

Section 11: Insider Activities

Venue and source note. Hanbell is a Shenzhen A-share. The primary insider-disclosure sources are the company's announcements on the Shenzhen Stock Exchange / cninfo (巨潮) - specifically pre-disclosure announcements of director/executive ("董监高") and major-shareholder share reductions (减持) and increases (增持). I attempted to retrieve the detailed transaction-by-transaction record over the last 12 months; the structured aggregator pages (Eastmoney's 高管持股变动 table) returned empty/unloaded data within the search budget, and the granular per-filing announcements were not fully retrievable. What I could establish from primary filing references is below, stated with that limitation.

What I could establish:

• A director/senior-management share-reduction pre-disclosure exists in the record: a planned reduction of up to 6,444,158 shares (about 1% of total shares) to be sold on the Shenzhen exchange over a defined window (a routine pre-disclosed director/executive reduction). (Hanbell director/executive share-reduction pre-disclosure announcement, Shenzhen exchange filing)
• A prior director-and-executive reduction plan was disclosed earlier (a February 2023 pre-disclosure announcement of partial reductions by certain directors and senior managers), indicating reductions by management are a recurring feature rather than a one-off. (Hanbell "关于部分董事和高级管理人员减持股份的预披露公告," 22 Feb 2023)
• Ownership structure as the backdrop: controlling shareholder Panama Hermes Company (~32.7%), ultimate controller Liao Zhemian (~7.55%), top-ten holders ~66.6% of float (as of 30 Sep 2025) - a tightly held register where management/family decisions dominate the stock's free supply.

Buys. I found no evidence of open-market insider buying by directors, executives or the controlling shareholder over the last 12 months within the search budget. There is therefore no cluster-buy or single-large-purchase conviction signal to flag.

Sells. The identified insider activity points toward periodic, pre-disclosed director/executive reductions (the ~1%-of-shares plan, consistent with the earlier 2023 pattern). For these, the disclosed rationale is the standard "personal funding needs" framing typical of Chinese director/executive reduction filings; a specific business-outlook reason is not disclosed and should not be inferred. These are pre-announced, capped, regulator-governed reductions rather than emergency selling.

Net assessment. On the evidence I could verify, insiders are net sellers in a routine, pre-disclosed fashion, with no offsetting open-market buying - a mild negative-to-neutral signal driven by recurring director/executive reduction plans rather than any single alarming transaction. The bigger governance feature is the concentrated, family-controlled register. I want to be explicit about the limitation: a complete, transaction-by-transaction 12-month insider ledger for this Shenzhen venue could not be fully retrieved within the search budget, so this section reflects the filings I could confirm rather than an exhaustive list. For a definitive record, the cninfo (巨潮资讯) announcement archive for 002158 is the authoritative primary source.

Section 12: Scenarios

Bull case. The solar trough proves to be the bottom, and 2025 is remembered as the low. Through 2026-2027 the semiconductor vacuum-pump business does what management said it would: the SEMI-certified PMF/iPM/iPH families move from first qualifications to repeat orders across multiple domestic fabs, revenue roughly doubles off its small base and keeps compounding as China's chip-capacity build-out absorbs domestic pumps in place of imports. At the same time the data-center cooling line catches the AI infrastructure wave - magnetic-levitation and air-bearing centrifugal compressors win designs into Chinese hyperscale and liquid-cooling deployments, turning the most advanced (and highest-margin) part of the compressor catalogue into a genuine growth engine. Solar stabilises and even partially recovers, so the vacuum line gets a second leg. The after-market keeps compounding under all of it. Hanbell stops being "the solar vacuum company that got hit" and becomes a diversified compressor-and-vacuum platform riding two structural localisation themes at once, with earnings recovering faster than revenue because of operating leverage.

Base case. Management delivers roughly what it has guided, with friction. Semiconductor vacuum grows quickly in percentage terms but stays a modest share of the whole for a couple more years - real, visible, but not yet the dominant driver. Data-center cooling compressors grow steadily as one of several compressor demand sources alongside cold chain, heat pumps and replacement demand, with competition keeping pricing honest. Solar remains soft and is a slow, uneven recovery rather than a snap-back, so the vacuum line takes time to return to former levels. The compressor business continues to carry the company as it did through 2025, the dividend stays linked to earnings, and profit grinds back up off the 2025 trough (as the Q1 2026 return to growth already hints) without a dramatic re-rating. A diversified, cyclically-recovering industrial with two credible options that are maturing on a multi-year clock.

Bear case. The two pivots disappoint relative to the solar hole. Semiconductor qualification proves slower and stickier than hoped - Edwards and Ebara defend their installed base, domestic fab capex wobbles, and the "doubling" guidance turns out to have been optimistic off a base too small to matter to group earnings. Data-center cooling becomes a margin fight as Danfoss, the global majors and a crowd of Chinese entrants converge on magnetic-bearing centrifugals, so Hanbell wins volume but not the pricing it expected. Meanwhile solar stays depressed for longer, leaving the once-rich vacuum segment structurally smaller. Earnings stay range-bound at the lower post-2025 level, operating leverage works against any further demand dip, FX losses recur on the export book, and recurring insider reductions cap sentiment. Nothing breaks catastrophically - the compressor base and after-market keep the company solidly profitable - but the growth story that justified the franchise quietly stalls, and Hanbell settles into being a steady, cyclical, single-digit-growth Chinese equipment maker rather than a localisation winner.

Sources

• Shanghai Hanbell company "About Us" (hanbell.xyz)
• Hanbell product / made-in-china profile
• DCFmodeling — Hanbell history, ownership, how it makes money
• FY2025 results review — StockStar (营收净利双降, 真空-63%, FX losses)
• Sina — FY2025 annual report breakdown (net profit -45.66%, cash flow +595%)
• Yicai — FY2025 net profit -45.66%, dividend RMB 4.5/10 shares
• Sina dividend history table (002158)
• 12 May 2026 annual online results briefing IR record (CFi.CN)
• 24 Apr–29 May 2026 IR record (CFi.CN)
• 7–15 Jan 2026 IR record (CFi.CN)
• 17–31 Dec 2025 IR record (CFi.CN)
• 23 Sep 2025 IR record (cninfo PDF)
• Q1 2026 result — STCN (net profit RMB 1.22bn, +3.3%)
• Hexun — Q3 2025 review, semiconductor vacuum + data-center thesis
• Zhihu deep report — Hanbell as domestic semiconductor vacuum-pump leader (70% crystal-pulling share)
• Vacuum-pump brand ranking (Leybold/Edwards/Pfeiffer/Ebara/ULVAC/Hanbell)
• Everbright Securities — FY2025/Q1 2026 review, data-center cooling outlook
• Dry vacuum pump market sizing report
• SEMI member directory — Hanbell Precise Machinery
• Hanbell iPH-series dry pump (distributor spec)
• Hanbell director/executive reduction pre-disclosure precedent (2023-02, Shanghai Securities News)

Section 13 (SemiAnalysis / Stratechery / MBI Deep Dives) is omitted: a real search of all three found no qualifying in-depth coverage of Hanbell.

A few honest caveats on completeness: the five management records are China's IR-activity-record equivalents of concalls (no quarterly earnings calls exist for A-shares); the most recent is within 90 days (12 May 2026). The granular 12-month insider ledger for this Shenzhen venue could not be fully retrieved within the search budget - Section 11 states what I could confirm and points to cninfo as the authoritative primary source. Per the report rules, no revenue/margin/valuation figures appear in the narrative; segment mix percentages and per-share dividends are used only where the brief permits.