| Meta Information | Content |
|---|---|
| Rating | Neutral |
| Target Price | HKD 19.8-23.1 (base-case fair value range, 12 months) |
| Current Price | HKD 18.55 (2026-10-12 close, a 52-week low; 52-week high HKD 55.65, drawdown of 66.7%) |
| Margin of Safety | Approx. +6.7% (base-case fair value floor HKD 19.8 vs current price, per system standard) |
| Probability-Weighted Expected Value | Approx. HKD 20.95 (approx. +12.9% vs current price; bear/base/bull = 30%/50%/20%) |
| Time Horizon | 12 months |
| Market Cap / Shares | HKD 8.16 billion / 440.0 million shares (393.8 million H shares + 46.2 million domestic shares) |
Dobot is a leading global cobot shipment vendor (per CIC Research, ranked first globally in 2025 with a 13.2% share, though third-party MIR/GGII data show discrepancies), currently operating a dual-business structure of "cobot base + embodied intelligence second growth curve": 2026H1 revenue was RMB 316 million, up 106.6% YoY, of which embodied intelligence revenue was RMB 45.23 million (14.3% of total; prior-year base only about RMB 2.10 million). The company has a thick balance sheet — cash-like assets of RMB 2.093 billion at end-June 2026, net cash of about RMB 1.95 billion (28% of market cap, roughly HKD 5.20/share) — but cash burn is accelerating: H1 operating cash flow was -RMB 259 million (vs -RMB 65 million a year earlier), inventory rose 67.8% in six months, USD deposits of about RMB 1.58 billion remain largely unhedged, and H1 FX losses were RMB 66.10 million. After a 66.7% share-price drawdown over the past year, valuation has been substantially reset from the early-2025 humanoid narrative bubble (PB 2.77x, EV/S(2026E) about 5-6x). The current price roughly sits in a "reasonably cheap" zone of "fairly priced cobot base + moderately priced embodied option," but the discount to this report's base-case fair value range (HKD 19.8-23.1) is only about 6.7% — a thin margin of safety. Key constraints: embodied intelligence order book (over RMB 60 million) contains a substantial proportion of framework/letter-of-intent agreements; the controlling shareholders' no-sale commitment expires on 2026-12-24; and the A-share listing, which passed review three months ago, still has not submitted its registration application. Conclusion: Reasonably cheap, watch and wait for validation — with A-share registration progress, Q3 operating data, and embodied order conversion as triggers to add; if price breaks below the net cash + base business support zone (about HKD 12-14), the thesis needs to be re-examined.
Opportunity positioning (evidence grade B): An event-driven opportunity at 52-week lows. The current price has pulled back to the "reasonably cheap" zone of "fairly priced cobot base + moderately priced embodied option," with right-skewed odds (approx. -40% to bear-case midpoint, approx. +85% to bull-case midpoint), but the margin of safety is thin — rated "watch" before validation; upgrades upon validation or when price enters the support zone of net cash (approx. HKD 5.2/share) plus standalone cobot valuation (approx. HKD 12.0-14.2/share).
Market environment (regime: neutral to headwind): The Hong Kong robotics sector is in a narrative-fade phase — Unitree hit a post-IPO low on 2026-10-08, and JPMorgan initiated coverage with an "Underweight"; as a high-beta small-cap loss-maker, Dobot has fallen harder than peers in this sector-wide de-rating. In such an environment, "cheapness" alone is no reason to rise; event validation is required.
Three factual changes most likely to move the stock over the next 6-12 months (ranked by impact):
Primary price drivers: A-share registration progress (re-rating anchor + financing leg) and embodied order conversion (the current price embeds an option premium of about HKD 4.3-6.6/share, 23-35% of market cap).
Expectations gap (status: mixed; sell-side anchors scarce): The only effective sell-side anchors are CICC at HKD 40 (2026-08-25, cut 34% from HKD 61) and Daiwa at HKD 62 (2026-04, not updated for 6 months); Futu's "consensus target of 40" is effectively single-source replication — market pricing itself may be distorted. Reverse-engineering the price shows the current level implies a 2026-2029 revenue CAGR of about 23% + breakeven by 2028 + embodied orders not slipping — modestly conservative versus the company's historical 31% CAGR and H1 doubling momentum. The real debate centers on two points: the A-share registration timing (bearishness already priced in; re-rating window not yet open) and the pace of embodied AI delivery (revenue is materializing but adjusted losses widened ~179% YoY, and market patience is wearing). Validation catalysts and falsification conditions: see the tracking framework in Chapter 12 (Q3 operating data, registration announcements, commitment expiry, gross-margin floor).
Interpretation and balance: Volume expansion is real — six-axis capacity utilization rose from 61.3% (FY2023) to 92.6% (FY2025); industrial scenarios (automotive/3C/equipment manufacturing) have overtaken education as the largest revenue source; customer concentration is low and declining (top-five customers at 10.9%). But three constraints must be seen simultaneously: ① Low base: revenue shows extreme back-loaded seasonality (2025H1 was only 31% of the full year; H2 was 2.22x H1), so a substantial part of the 2026H1 +106.6% comes from a low base and full-year growth will inevitably decelerate; ② Price erosion continues: the company's own A-share inquiry-response forecast shows six-axis volume CAGR falling to ~43% (vs 82% during the reporting period); ASP remains in a downtrend, with gross margin sustained only by cost reductions in harmonic reducers (2025 purchase price -26.7%) and a rising share of high-ASP embodied products; ③ Apparent gross-margin improvement is flattered: 2026H1 gross margin including inventory write-downs was 47.4% (+0.4pp YoY); excluding inventory write-downs, 48.9% vs 49.4% a year earlier is actually -0.5pp — volume expansion has not yet delivered structural improvement in unit economics. Market-share data must also be read side by side: CIC's "global No. 1" includes desktop/education units and is measured in shipments; MIR's data show JAKA as China's 2024 shipment leader (21.9%), and GGII's 2025 China top five does not include Dobot. This report positions its base business as "a global shipment leader, upper second tier in China."
Interpretation and balance: The second curve going from zero to 14.3% of revenue is the most important fundamental change of 2026; the third batch of full-size Atom humanoids entered mass delivery in 2026-02, with price率先 dropping below RMB 200k. But this report remains restrained on this thesis: ① the RMB 45.23 million has no product-line breakdown (Atom vs multi-legged vs service), and the timing gap between revenue recognition and shipment figures (media reported H1 shipments exceeding RMB 40 million) is invisible; ② the order book covers the ~RMB 55 million to be recognized in H2 only ~1.1x; the ~RMB 100 million full-year guidance (per inquiry response/media paraphrase; not directly disclosed in the original announcement) has no buffer, with zero disclosure on framework-contract conversion rates or penalty clauses; ③ on valuation, the market is paying about HKD 4.3-6.6/share (approx. HKD 1.9-2.9 billion) for the option — roughly 16-25x guided revenue. With Unitree's PB still near 60x and JPMorgan slapping an "Underweight" on the sector within a compression channel, this pricing is neutral — hardly a mispricing to exploit.
Interpretation and balance: Net cash of approx. RMB 1.95 billion (approx. HKD 2.29 billion, HKD 5.20/share) provides a hard 28% cushion at the current price, with a cash runway of roughly 3-4 years at annualized H1 operating cash flow — an essential distinction from typical cash-burning robotics startups. But the direction is deteriorating: H1 operating cash outflow was 2.4x the book loss (inventory up a net RMB 147 million, prepayments and receivables up RMB 43 million), staff costs RMB 157 million in six months (+146%), and bank borrowings +40% in six months; if H2 consumption holds flat, the actual runway will be shorter than a linear annualization. FX exposure must be stated precisely: the company explicitly has no systematic hedging policy, holding only a notional USD 30 million forward (coverage ~13%, matured and settled in 2026-07); H1 FX losses of RMB 66.10 million were about 4.2% of USD deposits — the exposure is structural, and the company's "temporary loss" characterization does not square with this. Also beware financing inertia: within 22 months of listing, the company completed an IPO plus two placements (at HKD 54.3/46.8, net proceeds HKD 1.793 billion), cumulatively diluting ~20%; placees in the two placements are now sitting on 60-66% paper losses and are unlocked financial investors; if the RMB 1.2 billion A-share raise (passed review three months ago, registration unfiled) drags past the 2027-07 validity window, the financing leg breaks and the company would most likely repeat the discounted-placement cycle.
Interpretation and balance: From the 2025 bubble (peak PB 83.8x; 2026E P/S of 27.5x at CICC's initiation) to PS(2026E) of ~7.3x now, the humanoid narrative premium has been cleared by about three-quarters. The current price implies — 2026-2029 revenue CAGR of ~23%, breakeven by 2028, embodied orders not slipping — achievable-to-conservative versus the historical 31% CAGR and H1 doubling momentum. But note the denominator of this report's base-case fair value (HKD 19.8-23.1) is 2027E revenue of RMB 1.15-1.32 billion; if embodied order conversion disappoints and 2027E is only RMB 1.0 billion, fair value falls to approx. HKD 17.2-19.8 at the same multiples (-7% to +7% vs current price) — the denominator debate is the largest valuation uncertainty in this case. Supply side: 193 million shares (~44% of share capital) unlocked in 2025-12; controlling shareholders Liu Peichao/Lang Xulin's no-sale commitment expires 2026-12-24 (Liu directly holds 21.35%, worth approx. HKD 1.74 billion at the current price, versus daily turnover of only about HKD 30-65 million); A-share dilution of ≤11.11% in the pipeline — any sale signal could break through the HKD 200 million incentive purchase plan's support. Overall judgment: reasonably cheap, with the current price at a ~6.7% discount to the base-case fair value floor and ~13% to the probability-weighted midpoint; odds are right-skewed (+15.6% to base case / +84.9% to bull case, -39.6% to bear case), but the thin margin plus supply overhang do not support a higher rating.
(Unit: RMB hundred million; fiscal year ending 31 December; losses shown with negative signs)
| Metric | FY2023 | FY2024 | FY2025 | 2026H1 (latest period) |
|---|---|---|---|---|
| Revenue | 2.87 | 3.74 | 4.92 | 3.16 (+106.6%) |
| Gross margin | — (unverified) | 46.6% | 46.1% | 47.4% (48.9% excl. inventory write-downs) |
| Net profit attributable to shareholders | -1.03 | -0.95 | -0.84 | -1.06 (loss widened 158.7%) |
| Adjusted net loss (non-IFRS) | — | -0.36 | -0.50 | -0.64 (vs -0.23 a year earlier) |
| Net margin | -36.0% | -25.5% | -17.0% | -33.4% |
| Operating cash flow | -1.58 | -0.92 | -0.43 | -2.59 |
| Free cash flow | — | — | -0.57 | -3.23 |
| Cash-like assets (cash + time deposits + structured deposits) | — | — | ~24.3 (year-end; cash 5.80 + time deposits 18.45) | 20.93 (2026-06-30) |
| Interest-bearing liabilities | — | — | ~1.1 (bank loans 0.72 + leases) | 1.42 (2026-06-30) |
| Net cash | — | — | — | 19.51 (approx. HKD 5.20/share) |
| Liabilities/assets ratio | — | — | — | — (net assets approx. RMB 2.55 billion; total assets not separately extracted) |
| Six-axis shipments (10k units) / ASP (RMB 10k/unit) | 0.25 / 5.66 | 0.46 / 4.72 | 0.83 / 3.82 | Not disclosed (embodied ASP RMB 89.1k/unit) |
| Contract liabilities | — | — | 0.11 (year-end) | 0.22 (+103.1% vs year-end) |
Drivers of metric changes (items with YoY ≥ ±20%, with company explanations attached): ① Revenue +106.6% — company attributes to six-axis cobots +110.4% and embodied intelligence +2,052.4%; ② R&D expenses +148.4% to RMB 102 million — "increased staff costs and share-based payment expenses, and increased R&D materials input" (embodied AI ramp-up); ③ Selling expenses +57.7%, administrative expenses +98.5% — globalization and headcount expansion; ④ Other expenses RMB 1.80 million → RMB 67.90 million — "increased FX losses on foreign currency holdings" of approx. RMB 66.10 million + RMB 1 million in charitable donations; ⑤ Operating cash flow -RMB 64.9 million → -RMB 259.2 million — the company did not separately explain the cash flow deterioration, only indirectly citing "increased R&D investment and temporary FX losses"; at the cash-flow-statement level the main drivers were a RMB 147 million inventory increase and a RMB 43 million increase in prepayments and other receivables; ⑥ Capex RMB 4.10 million → RMB 63.80 million — new office fit-out plus office and computing equipment purchases; ⑦ Inventory +67.8% vs year-end — no company explanation disclosed; ⑧ Share-based payment expenses +132% to RMB 41.59 million — expanded staff incentives (in 2026-02, an additional 8.638 million options granted at HKD 37.78 plus 1.531 million RSUs at HKD 1.0).
2026H1 revenue of RMB 316 million landed at the midpoint of the company's July voluntary disclosure range (RMB 300-330 million); net loss attributable to shareholders of RMB 105.7 million, at the deeper end of the guidance range (-RMB 90-120 million); roughly breakeven excluding FX losses and share-based payments (this report calculates approx. +RMB 2 million on an IFRS basis; the company's A-share basis self-reported adjusted losses in the -RMB 35-65 million range). Structure matters more than headline: industrial scenarios +142.4%, 62.8% of revenue (education down 14pp to 23.4%); overseas revenue +90.9%, 48.4% of total — the base business's "industrialization + globalization" shift is materializing. Negatives in three areas: operating cash flow of -RMB 259 million (vs -RMB 65 million a year earlier), inventory +67.8% in six months, and FX losses of RMB 66.10 million. Market reaction skewed negative: the day after the interim results, CICC cut its target price from HKD 61 to HKD 40 (-34%); the sector has accelerated its decline since late August, and the stock hit a 52-week low on 10-12 — the combination of growing revenue with growing losses and accelerating burn has overwhelmed the positive reading of doubled revenue. Versus sell-side expectations: coverage is thin (only CICC and Daiwa), with no robust consensus data; this report does not render an "above/below expectations" verdict on the market's behalf, only benchmarking against the company's own guidance ranges (met) and management's 2026 embodied revenue guidance (45% completed in H1).
Business model in brief: An asset-light hardware manufacturer and product-driven company with high R&D intensity. Fixed assets account for only 7.0% of total assets, and outsourced processing fees represent roughly 2.3% of costs; revenue comes from one-off equipment sales (service revenue is only about 1.0%, with no subscription/recurring revenue). With an in-house core component self-development rate above 90% (controllers/integrated joints/safety skins), the company internalizes upstream gross margin; its 2026H1 gross margin of 47.4% is significantly higher than Estun's (about 26%) and UBTech's (37.7%) — yet pricing power is limited (six-axis ASP has declined for three consecutive years); the premium comes from cost-reduction capability and product mix, not price increases.
Cash content of earnings (measured during the loss-making period): OCF/book loss improved from FY2023 1.53 → FY2024 0.96 → FY2025 0.51, then jumped back to 2.40 in 2026H1 (burning 2.4 yuan of cash per 1 yuan of book loss); FCF/loss reached 2.99 in 2026H1. Net working capital absorption was RMB 203 million against a revenue increment of RMB 163 million — about RMB 1.25 of working capital tied up per 1 yuan of incremental revenue, indicating an inventory build-up expansion phase. Recurring profitability: excluding SBC (RMB 41.59 million), listing expenses, and FX losses (RMB 66.10 million), 2026H1 was roughly +RMB 2 million, meaning operations are already near breakeven — but this relies on the assumption that "FX losses can be excluded," while the USD exposure is structural; meanwhile, the company's Non-GAAP measure adds back SBC in full every year (39.3% of the attributable loss), a loose definition that systematically understates labor costs in the adjusted loss.
Return on capital: ROIC is negative during the loss-making period; invested capital (net assets of about RMB 2.55 billion plus interest-bearing debt of RMB 140 million) is still in the cash-burning moat-building stage and cannot be benchmarked against WACC — validation of a true moat must wait until profitability turns around.
CapEx quality: FY2025 CapEx/depreciation was only 0.42 (historically under-invested, asset-light); it jumped to 2.93 in 2026H1 (RMB 63.8 million, new offices + computing equipment), which is expansionary; the company has not disclosed a maintenance vs. growth split. Capacity utilization of 92.6% (six-axis) shows existing lines are near full load, making expansion necessary.
Moat and red flags: Moat = integrated in-house components (gross margin 11–21pp above peers) + full 0.25–20kg payload matrix + channels in 100+ countries + repurchase rate above 60%. Red flags: ① price-for-volume coexisting with inventory build-up (ASP -32.5% over two years while inventory +67.8%); ② highly concentrated USD deposits with essentially no hedging; ③ a Non-GAAP measure that adds back large, recurring SBC in full.
Words-vs-deeds consistency: partially delivered, leaning pragmatic. ① Voluntary disclosure of H1 guidance range in July 2026 (revenue RMB 300–330 million; attributable loss RMB -90 to -120 million) — actual revenue of RMB 316 million and attributable loss of RMB -106 million both fell within the range (delivered); ② IPO proceeds of HKD 204 million earmarked for "production line and manufacturing capability development" — only 15.1% utilized as of June 2026, clearly behind schedule (not delivered; funds shifted to time deposits/wealth management products); ③ FY2025 annual report claimed "annual operating targets fully achieved" — revenue +31.7%, loss narrowed 11.9%, corroborated by CIC third-party share data (essentially delivered).
Shareholder friendliness: financing-oriented, neutral-to-negative bias. No dividends and no cancellative buybacks since listing; within 22 months, IPO plus two high-priced placements (HKD 54.3/46.8) raised a net HKD 1.793 billion with cumulative dilution of about 20%; an A-share issuance of up to 11.11% is in the pipeline; the current HKD 200 million "share purchase" is in fact trustee funding purchases for equity incentives (800,000 shares bought at an average price of about HKD 25.75), not a discretionary buyback. Positives: Liu Peichao and Lang Xulin voluntarily committed from December 25, 2025 to a 12-month no-sale pledge; RSU grant price of HKD 1.0/share (market price HKD 35.92) provides strong binding of core employees.
Risk signals: ① An unsecured related-party loan of RMB 18.63 million to Chairman Liu Peichao (to cover personal income tax on shareholding conversion, LPR-based interest, fully repaid by December 2025) — small in amount but sensitive in nature; ② Dispute with former employee Song over a shareholding platform: arbitration filed and a real-name whistleblower complaint lodged on July 17, 2026 (media estimates the disputed shares at about 5,957,600 shares, or 1.35% of total share capital); the company passed the A-share review the same day and later issued a clarification stating ownership is clear — the unresolved arbitration is a potential variable in the A-share registration process; ③ During the reporting period, executive director Wang Yong resigned and the joint company secretary changed — turnover at the governance level.
By application scenario (2026H1, % of product sales revenue):
| Scenario | Revenue share | YoY | One-line business logic |
|---|---|---|---|
| Industrial | 62.8% | +142.4% | Material handling/assembly/inspection for automotive/3C/equipment manufacturing; largest growth engine |
| Education | 23.4% | +26.9% | K12/vocational/higher-ed traditional base; share down about 14pp YoY |
| Commercial | 8.1% | +84.1% | Retail/health therapy; humanoid robot landing scenarios |
By product line (2026H1, % of total revenue): Six-axis collaborative robots 62.2% (+110.4%); embodied intelligence robots (humanoid + multi-legged) 14.3% (+2,052.4%, ASP of RMB 89,100/unit — the highest-value product line); four-axis 13.2% (+5.1%, strategic shift in education from four-axis to six-axis); composite robots 8.7% (+89.7%, project-based palletizing/welding integration). By region: Domestic 51.6% (+123.9%, direct sales share rising to 57.9%, relatively lower gross margin); overseas including HK/Macau/Taiwan 48.4% (+90.9%, localization in Europe/Americas/APAC, relatively higher gross margin).
Profit drivers and structural interpretation: The company does not disclose gross margin by segment, so the main profit contributor cannot be computed directly; but it can be inferred — industrial six-axis wins on volume (ASP RMB 38,200, gross margin dragged by domestic direct sales), education four-axis/six-axis has historically been the high-margin base, and embodied intelligence, with RMB 89,100 ASP × premium on new products with a high share, contributes the highest unit economics of incremental revenue. Evidence for gross margin structure differences exceeding 10pp across scenarios or product lines (neither disclosed) is insufficient; only a clear directional gap exists between domestic vs. overseas (the company itself states domestic gross margin is relatively lower). Company-specific segment risk: Education's share has fallen from roughly half of revenue pre-listing to 23.4% — the revenue mix completed its "education → industrial" switch within one year, an extremely fast pace; this can be read positively as a successful base transformation, but there is also the concern that slowing education revenue growth (+26.9%) drags on overall gross margin.
Accounting red flags (3):
Cross-period consistency (5 items, 5 reporting half-years assessable):
Current market data (October 12, 2026 close): Share price HKD 18.55; market cap HKD 8.16 billion (about RMB 6.96 billion); PS(TTM) 10.6x (TTM revenue RMB 655 million); PB 2.77x — lowest percentile since listing, but the sample spans only 22 months including the extreme 2025 humanoid bubble period, and net assets were inflated by two high-priced placements with negative ROE; the percentile is for reference only, not a cheapness/expensiveness headline. EV/S(2026E) of about 4.9–5.7x (on our own revenue estimate of RMB 890 million–1.02 billion; on the company's July-implied ~RMB 700 million basis, about 7.2x) — this denominator dispute directly determines whether the valuation is "in line with UBTech" or "above UBTech."
| Peer | Market cap | Revenue growth | Gross margin | ROE | Key differentiator |
|---|---|---|---|---|---|
| Dobot (02432.HK) | HKD 8.16 bn | 2026H1 +106.6% | 47.4% | Negative (loss) | Top global cobot shipments + embodied pivot; thickest cash cushion |
| UBTech (9880.HK) | HKD 34.08 bn | 2026H1 +104.2% | 37.7% | Negative | Humanoid revenue already over 41% of mix, 4x the scale, customer concentration |
| Estun (002747.SZ) | RMB 25.20 bn | 2026H1 +1.1% | ~26% | ~0.5% | No.1 domestic traditional industrial robot player; "big but thin" |
| Efort (688165.SH) | RMB 7.34 bn | 2026H1 +28.9% | 15.3% | Negative | Price-war casualty sample; cautionary tale of collapsing gross margin |
| Inovance (300124.SZ) | RMB 136.71 bn | 2025 +21.8% | 29.0% | 16.3% | Profitable industrial automation platform anchor; a giant potentially moving into robots |
What the market has already priced in: EV after deducting net cash is about RMB 5.01 billion. Solving backwards: EV ÷ 30x forward PE ≈ implied 2029 attributable net profit of about RMB 167 million, corresponding to revenue of about RMB 1.76 billion at a net margin of about 9.5% — i.e., the current price requires the company to deliver "2026–2029 revenue CAGR of about 23% + breakeven in 2028 + net margin climbing to nearly 10% by 2029," while the reality is H1 revenue +106.6% with adjusted losses still widening. Sum-of-parts residual: valuing the cobot business standalone (2026E ~RMB 850 million × 3.0–4.0x PS + net cash) gives about HKD 12.0–14.2/share; the current price is HKD 4.3–6.6 above that — this is the portion the market pays for the embodied/humanoid option (23–35% of market cap, about 16–25x guidance-basis revenue).
Three-layer value (EPV): Asset value (floor) = net cash of about HKD 5.20/share; EPV (zero growth) ≈ 0 — 2026H1 recurring earnings excluding SBC and FX are roughly breakeven, and normalized FCFE is negative during the expense expansion phase (this report's EPV inputs: basis=equity, normalized cash flow per share 0, cost of equity 11.5%); growth option ≈ nearly 100% of the current price. This is not a "market paying only for zero growth" profile — for a loss-making growth stock, the price being almost entirely supported by the growth option is the norm; here, the floor consists of HKD 5.20/share net cash plus a HKD 12–14 cobot base, and the option component is concentrated in the embodied second curve, whose realization depends on training-policy orders, Optimus supply-chain spillover, and embodied revenue stepping up in 2027.
Three scenarios and odds (12 months; fair value ranges are derived from independent assumptions rather than anchored to the current price; the base-case exit multiple EV/S 4.5–5.5x is anchored between Estun's TTM PS of 5.1x (profitable leader floor) and UBTech's 10.98x (humanoid ceiling), independent of the company's current price):
| Scenario | Probability | Fair range | vs. current | Key drivers |
|---|---|---|---|---|
| Bear | 30% | 10.6–11.8 | -36%~-43% | Embodied thesis falsified (framework orders not converting) + price war continues (2026E revenue at the ~RMB 890 million low end, 2027 stall) + net cash down to ~RMB 1.75 bn after two more years of cash burn, cleared at PS 2.5–3.0x |
| Base | 50% | 19.8–23.1 | +7%~+25% | Cobots +25–30%, embodied delivering RMB 100 million, 2027E revenue RMB 1.15–1.32 bn × EV/S 4.5–5.5x + net cash RMB 1.95 bn; if 2027E is only RMB 1.0 bn, fair value drops to about 17.2–19.8 |
| Bull | 20% | 32.4–36.2 | +75%~+95% | Embodied scaling to RMB 400–500 million in 2027 (10,000-unit training orders + supply-chain spillover), A-share listing re-rating, 2027E ~RMB 1.45 bn × EV/S 7–8x (below UBTech) + ~RMB 2 bn net cash, already deducting 10–15% A-share dilution |
The current price of 18.55 sits at roughly the 31st percentile of the bear-bull distribution, slightly below the base-case floor — odds are right-skewed but not deep: probability-weighted expectation of about HKD 20.95 (+12.9%), downside to the bear-case midpoint at -39.6%, upside to the bull-case midpoint at +84.9%. External anchor comparison: CICC HKD 40 (above this report's bull-case floor; its 2030 13x P/S long-term anchor is aggressive); Daiwa HKD 62 (not updated since April 2026, timeliness questionable); no independent bearish sell-side target price exists for Dobot (coverage gap).
Our own earnings forecast (reconciliation anchor for the next 12 months):
| Period | Revenue | Net profit | Key drivers |
|---|---|---|---|
| FY2026E | RMB 850–950 million (midpoint 900) | Attributable -190 to -230 million; adjusted -120 to -150 million | Six-axis shipments +80% or more; embodied ~RMB 55 million recognized in H2 (~RMB 100 million full year); gross margin 46–48%; H2 expense growth decelerating |
| FY2027E | RMB 1.15–1.32 bn (midpoint 1.22 bn) | Attributable -60 to -120 million; adjusted -10 to -60 million | Cobots +25–30% (overseas + share gains); embodied RMB 250–350 million; revenue leverage declining, expense ratio down 5–8pp |
Cross-checked against management guidance (~RMB 100 million embodied in 2026; A-share prospectus targeting breakeven by 2028) and sell-side views (CICC implying a near-breakeven 2026): our revenue forecast is above the company's July-implied basis (~RMB 700–770 million) and below mechanical seasonal extrapolation (RMB 890 million–1.02 billion) — center-left, i.e., moderately conservative. Conclusion: whether this is undervalued hinges on the 2027 revenue denominator and embodied order conversion; the current price is reasonably cheap, with quality (gross margin, cash, share) stated separately from price (thin margin) — a good company, a price after clearance, and an option that is not yet cheap.
1. Industry Size: Dual-Track Overlay. Collaborative robots (core business): Per GGII, the global market reached RMB 10.11 billion in 2024 (RMB 3.61 billion in 2020, CAGR 29.4%), and approximately RMB 12.13 billion in 2025; the China market was RMB 3.19 billion in 2024 (+20.3%), with over 40,000 units sold in 2025. Structural penetration benchmark: Per IFR, cobots accounted for 10.5% of global new industrial robot installations in 2023 (~54,000 out of 541,000 units) — penetration remains low and the ceiling is far away. Interact Analysis projects a shipment CAGR of ~20% for 2024-2029 (2024 was a shipment trough, with recovery from 2025). Humanoid robots (second curve): Goldman Sachs' upgraded forecast (August 2026) shows global shipments of 75,000 units in 2026 → 890,000 units in 2030 → ~6.5 million units in 2035 (corresponding to a market of ~USD 138.3 billion, implying a 2026-2035 CAGR of ~64%); per GGII/East Money Securities, the China market grows from RMB 9.0 billion in 2026 → RMB 240 billion in 2030 (shipment CAGR 85%). Actual progress: global shipments in 2025 were ~18,000 units (+~500%), and 2026H1 already reached 19,100 units (+272%, 97% from Chinese manufacturers) — the year of mass production has arrived, but demand is concentrated in research/education and entertainment (Unitree derived 73.6% of revenue from research and education, IDC); real industrial paid demand will only scale up after the policy-driven practical training initiative at end-2026. The two tracks differ by an order of magnitude in size: collaborative robots are a tens-of-billions-RMB growth market with 20% growth; humanoid robots are a single-digit-billion-RMB option market with triple-digit growth.
2. Supply Chain and Value Distribution. Upstream core components (servos + harmonic reducers + controllers) account for 60-70% of total machine cost; gross margins by segment: reducers ~40% > servos ~35% > controllers ~25% — the fattest gross-margin link is precision transmission. Localization rates: the complete cobot segment has reached 90.25% (2024, per JAKA's prospectus citing MIR); harmonic reducer localization is ~30-40% (Leaderdrive is the domestic leader, with costs ~30% lower than Japanese rivals; Leaderdrive accounts for 13.5% of Dobot's top-five supplier purchases). Midstream complete machines: globally, UR (Teradyne) + the Big Four + the Chinese camp (JAKA/UBTech/Elite/Dobot/Rokae/FR). Downstream: ~70% of demand is in industrial scenarios (handling/machine tending/assembly/palletizing over 50%, GGII 2025 Blue Book); payback period for a single palletizing cobot is ~12 months. Dobot's position: midstream complete machines + upstream extension (self-developed core components exceed 90%, including controllers/integrated joints/safety skin), internalizing 35-40% of upstream component gross margins — this is the structural reason its gross margin exceeds Estun's by 21pp and UBTech's by 10pp; it has strong bargaining power over upstream (multiple suppliers + in-house R&D), but downstream is mainly distribution across 100+ countries with limited captive volume, and it faces the complete-machine price war head-on.
3. Supply-Demand and Competitive Landscape. Demand side: rising labor costs + room in robot density (China at 470 units per 10,000 workers vs. 1,000-unit level in South Korea/Singapore) + flexible production line upgrades in 3C/new energy/semiconductors; China's industrial robot output for Jan-Aug 2026 was 730,000 units (+29%), and IFR forecasts 655,000 new global installations in 2026 — volume is recovering while prices decline (average cobot prices in China down ~30% over two years). Supply side is consolidating: JAKA withdrew its STAR Market IPO in January 2026 (2025H1 net loss of RMB 19.97 million, growth slowing to +4.3%), and Teradyne's robotics division cut 14% of staff in 2025 with declining revenue; meanwhile, humanoid-track newcomers are exploding (over RMB 100 billion in embodied AI financing in China in 2026, with 140+ companies funded) — cobots are consolidating while humanoids are in an arms race, presenting Dobot with market-share transfer opportunity on one front and financing-burn pressure on the other. Concentration: cobot concentration is not high (China CR1 is JAKA at 21.9% per MIR 2024 sales; globally UR is No. 1 at ~15%, per aggregator data pending verification; per CIC, Dobot was global No. 1 in 2025 at 13.2%) — the discrepancy stems from Dobot's shipments including many desktop/education models. Humanoids: in 2026H1, AgiBot led globally with 8,400 units at 44% share; Unitree + AgiBot together exceeded 50% of global shipments in 2025. Entry barriers: hardware barriers have been leveled by China's supply chain (90%+ component localization); true barriers lie in motion control algorithms and safety certifications, global channel brands, application process packages, and scale-driven cost advantages.
4. Cycle and Regulation. Cycle position: collaborative robots = mid-growth stage of the lifecycle (10.5% penetration) + the upward recovery leg after bottoming in 2024; humanoids = the critical inflection from introduction stage to mass-production ramp, with valuation and financing far ahead of profitability. Leading indicators: China's monthly industrial robot output (+29% confirms upcycle), cobot ASP downtrend (unstopped = price war continues), leading humanoid makers' orders/deliveries (UBTech Walker S2 annualized capacity exceeds 6,000 units; Tesla's target rising from 1,000 to 2,000-2,500 units per week), 3C/new energy capex, and primary-market financing. Strong regulatory/policy catalysts: the "Robot+" Application Action (2023, density-doubling target), the Guidelines for Innovative Development of Humanoid Robots (2023, world-advanced by 2027), embodied AI consecutively written into the 2025/2026 Government Work Reports, MIIT + SASAC's 2026 Special Action on Real-Scenario Practical Training for Humanoid Robots and Embodied Intelligence (June 2026: 100 high-value scenarios by year-end, 10,000-unit-level deployment capability), and Shenzhen's Intelligent Robotics Industry Work Plan 2026-2028 (half-hour supply chain circle + Robotics Valley). Policy cuts both ways: training orders are demand, but subsidy-driven low-price competition is the enemy of industry gross margins.
5. Quantified Demand Inflection Chain. For Dobot, humanoids represent "second-curve category expansion" rather than a usage surge in existing products: driver trajectory (global humanoid shipments from 18,000 units in 2025 → 19,100 in 2026H1 → 890,000 by 2030) × surge in per-unit value (cobot RMB 38,200 → humanoid ~RMB 199,000, roughly 5x) × penetration ramp (research/education → 10,000-unit training deployment by end-2026 → industrial volume ramp in 2028-2030) × company share (currently <2%, hundreds of units — far behind Unitree's 5,500 / AgiBot's 8,400 / UBTech's 1,079 full-size units). Quantified conclusion: the industry-side chain is verifiable (Goldman Sachs/IDC/policy benchmarks), while only the company-side revenue guidance is verifiable (~RMB 100 million in 2026) — the 2026 P&L contribution remains small (~11-15% of total revenue); for Dobot, humanoids are an option rather than a current inflection point, and the cobot core business (~85% of revenue) is the source of earnings elasticity. This mutually corroborates the "option premium of 23-35% of market cap" pricing in the E6 valuation.
6. Company Industry Position. Global cobot shipment leadership: per CIC, global No. 2 in 2023 (13.0%) → global No. 1 in 2025 (13.2%, cumulative installations exceeding 100,000 units, No. 1 among Chinese players in overseas shipments for eight consecutive years); competing benchmarks: per MIR, JAKA was China's sales leader in 2024 (21.9%), and UR remained global No. 1; GGII's 2025 China top five (JAKA/UBTech/Rokae/Huasheng Control) does not include Dobot — differences arise from "shipments vs. installed value" and "including desktop education models vs. six-axis industrial" caliber gaps; this report does not adopt the "No. 1" narrative from any single benchmark. Share trend is slightly rising (13.0% → 13.2%) alongside UR's revenue decline and JAKA's slowdown — the share-transfer thesis is supported by peer facts. Competitive profile: volume up, price down (ASP -32.5% over two years), slight share gains + highest gross margin among peers (self-developed integration) + small revenue base (1/4 of UBTech's, 1/10 of Estun's) + still loss-making. Positioning conclusion: a global leader in the cobot track + a chaser in embodied AI transformation.
Generic financial reporting frameworks tend to miss four defining metrics for robotics product companies; this report tracks them separately:
① Shipments and Capacity Utilization: Six-axis shipments of 2,500 → 4,600 → 8,300 units (FY2023-2025), capacity utilization 61.3% → 78.5% → 92.6% — full capacity is the direct rationale for the A-share fundraising expansion; but shipments were not disclosed in 2026H1 (a deduction for disclosure quality); track whether Q3/annual reports resume disclosure.
② ASP and Per-Unit Gross Profit (this report's estimates): Six-axis ASP fell from RMB 56,600 → 47,200 → 38,200; at a 46% gross margin, per-unit gross profit approximates RMB 26,000 → 22,000 → 18,000 (~-32% over three years) — per-unit economics are shrinking, offset by doubling unit volume; embodied AI ASP of RMB 89,100/unit (2026H1) is the only category with rising prices (new-category premium); in the A-share inquiry response, CR-series ASP is projected to decline only 0.5-1.8% annually during the forecast period (the narrowing-decline assumption is an implicit precondition for breakeven in 2028 — track and verify).
③ Scenario Penetration and Repeat Purchases: Three scenarios — education (K12/higher ed) → industrial (automotive/3C/semiconductor) → commercial (retail/physiotherapy); industrial share rose from ~48% to 62.8% within a year; customer repeat purchase rate exceeds 60%, top-five customers account for only 10.9% — healthy dispersion; contract liabilities of RMB 22 million (+103% vs. year-end) are a leading order indicator. ④ Humanoid product cycle: Atom's pricing was cut below RMB 200,000 in March 2025 → third batch delivered in mass production in February 2026 → fourth batch pending delivery; the "one brain, multiple bodies" platform (self-developed WAM embodied foundation model, average completion rate of 99.25% across LIBERO's four suites — the company claims it surpasses public models like GR00T-N1.7) drives reuse of the same tech stack across four form factors: cobots/bipeds/wheeled/multi-legged — product iteration cadence (one batch per half year) is faster than peers, but each mass-production batch (hundreds of units) is insufficient to amortize R&D.
Tracking discipline: Among these four metrics, whether ASP decline narrows (CR series <2%/year) and whether embodied revenue share can rise from 14.3% to 20%+ are the two numbers over the next two reporting periods most capable of confirming or refuting this report's judgments.
Overall rating: Neutral (confidence 0.55, 12 months). Quality and price addressed separately — quality: a global shipment leader's core business + highest gross margin among peers + net cash of RMB 1.95 billion + genuine second-curve revenue, constituting "a real transformation with a cash cushion" rather than pure narrative; price: valuation has been deeply reset to reasonably low, but the discount to baseline fair value is only 6.7%, and the next 15 months combine supply overhang (lockup-expiry memory, no-sell commitment expiring 2026-12-24, A-share dilution) with catalyst suspension (registration on hold). Strategy: at current prices, neither chase shorts nor rush to load up on the rebound — the right posture for an event-driven opportunity is to wait for verification: A-share registration submission/approval, Q3 operating data confirming quarter-over-quarter embodied volume growth, or a single formal order ≥RMB 50 million — any one materializing justifies an upgrade; conversely, if 2026 full-year embodied revenue falls below RMB 70 million, gross margin ex-impairment drops below 46%, quasi-cash falls below RMB 1.2 billion, or a deeply discounted placement is launched, downgrade the thesis.
Risk and Counter-Evidence Tracking (Written Summary): ① Embodied order conversion (coverage only 1.1x, including framework agreements) — full-year 2026 below RMB 70 million would falsify C2; ② Price war (the cautionary tale of Efort's gross margin collapsing to 15%) — gross margin ex-impairment below 46% for two consecutive periods would falsify C1/C4; ③ Cash burn and FX — quasi-cash below RMB 1.2 billion, RMB appreciation causing FX losses on the order of RMB 100 million+, or a placement at >15% discount would falsify C3's "thick cash" judgment; ④ A-share registration — no registration by July 2027 would invalidate the H-to-A thesis; ⑤ Supply — if major shareholder sell-downs are disclosed after the commitment expires on 2026-12-24, the buyback floor would be breached. Upside tracking: A-share registration approval and offering pricing (A/H linked re-rating), Q3 embodied revenue share breaking 15%, the training consortium list and SOE orders, and expansion of sell-side coverage (currently only CICC/Daiwa). Liquidity note: daily turnover of ~HKD 30-65 million; institutional position-building requires tranches, and volatility on event days will be amplified by thin liquidity.
This report's price anchor is frozen at the 2026-10-12 close (HKD 18.55); financial data are as of the 2026-06-30 interim report (RMB basis); exchange rate: RMB 1 = HKD 1.17188 (2026-10-13). Items marked "this report's estimates" are our own calculations based on company-disclosed data, not company-disclosed figures.