ROKAE

1.1 Identity, legal footprint, and operating model

ROKAE (Shandong) Robotics Group Inc. designs, develops, manufactures, and commercializes intelligent robots and robotic solutions across industrial, flexible collaborative, and embodied-intelligence categories. The listed issuer's filing dates incorporation in Beijing to December 14, 2014, while one independent profile calls 2015 the founding year; this appears to be a legal-incorporation-versus-operating-origin convention rather than a resolved factual match. The company converted to a joint-stock company in December 2024 and now has its registered office in Shandong. Separately, its website identifies Beijing as the management and R&D headquarters, with R&D centers in Beijing, Tokyo, Shanghai, and Shenzhen. These location labels should not be collapsed into a single headquarters claim because they describe different corporate functions.[CO001, CO002, CO003, CO004, CO005]

1.2 Leadership, control, and governance

Founder Tuo Hua is the ultimate controlling shareholder and combines the roles of executive director, board chair, and general manager; the interim announcement also describes him as chief executive officer. Tang Shengtian is an executive director and chief financial officer. This gives ROKAE concentrated founder leadership, potentially supporting consistent execution but increasing succession and oversight dependence on one individual. The company expressly disclosed that combining chair and chief executive roles departs from Hong Kong Corporate Governance Code provision C.2.1, although its board believes the arrangement preserves the balance of power and says it will periodically review a split. The disclosed leadership table is intentionally partial because the assigned evidence does not provide complete biographies or an exhaustive senior-management roster.[CO006, CO007]

1.3 Capital history and public-market transition

ROKAE began trading on the Hong Kong Stock Exchange on July 9, 2026 under stock code 03752.HK after offering shares at HK$38.00. Following full exercise of the over-allotment option, the company reported about HK$1.01 billion of gross proceeds and HK$937.1 million of net proceeds; these are financing inflows, not revenue or cash available for unrestricted use. A niche independent publication reports approximately US$210 million raised across nine private rounds before the IPO, including a US$6 million Series A in 2017 and a pre-IPO round above RMB500 million in April 2024. Those private-round totals were not reconciled to primary transaction documents in the assigned evidence, and no supportable current equity valuation was found. The stakeholder map therefore separates verified public-offering facts from reported private-financing history.[CO008, CO009, CO015]

1.4 Scale, financial snapshot, and diligence posture

The public record now provides useful dated scale markers. ROKAE reports serving more than 1,000 enterprises in over 40 countries and regions, and its June 2026 interim filing reports 807 employees. For the six months ended June 30, 2026, revenue was RMB414.7 million and the accounting loss attributable to owners was RMB79.0 million; revenue is not ARR and the accounting loss is not cash burn. The company also claims annual factory capacity of 50,000 robots and more than 600 patents and technology awards, but those figures lack independent verification in this pool. Evidence on adverse events is incomplete: the filing reported no material litigation likely to have a material adverse effect, while independent commentary flags continued losses, weaker overseas mix, and segment concentration. That combination supports monitoring, not a conclusion that no incidents exist.[CO010, CO011, CO012, CO013, CO014, CO016]

1.5 Exhibits

2.1 Market boundary and substitutes

The assigned evidence supports a robotics-market boundary centered on robot hardware, control systems, deployment services, and adjacent embodied-intelligence platforms. Across the company's charter, brand pages, and investor materials, ROKAE consistently presents three core product families: industrial robots, flexible collaborative robots, and embodied intelligent robots, plus robotic solutions and ecosystem tooling. That makes the relevant market narrower than generic "AI" and narrower than all factory automation spend: the included spend is the robotic body, control stack, integration-ready software, and related deployment services used in manufacturing, healthcare, commercial-service, and embodied-AI workflows. The practical substitutes are traditional industrial automation, competing collaborative robots, custom robotic solutions, and manual workflows in precision assembly or welding rather than general-purpose SaaS budgets.[CM001, CM002, CM003, CM004, CM005]

2.2 Sizing lenses and current monetization

Public market sizing in the assigned pool is directional rather than decision-grade. The clearest independent lens says China's multi-joint robot market could expand from RMB 21.9 billion in 2024 to RMB 45.3 billion by 2029, with the collaborative sub-segment growing faster than the broader category. That broad lens fits ROKAE's current mix because the company is already monetizing industrial, collaborative, and embodied-intelligent robots, and the latest HKEX filing shows embodied robots reaching 34% of first-half 2026 revenue. What the pool does not provide is a clean bridge from that large category TAM to the narrower portion ROKAE can actually serve by geography, industry, and budget owner, so the chapter can frame the opportunity but cannot responsibly publish a bottom-up SAM or SOM estimate.[CM006, CM007, CM010, CM011, CM015]

2.3 Buyer segments and adoption path

The visible adoption path starts with industrial users and their integration partners rather than direct consumer demand. Company-controlled pages show live focus sectors in auto parts, electronics, metal fabrication, healthcare, commercial services, and general industry, while the EU and investor-facing materials describe use cases such as flexible manufacturing, precision assembly, welding, and humanoid-arm R&D. The company also emphasizes open control platforms, AI interfaces, and SDK documents, which suggests a market-access strategy built around developers, system integrators, and ecosystem partners who can shorten deployment cycles. Public evidence is weaker on who exactly owns the budget inside each customer account, so sector and workflow visibility is stronger than procurement visibility.[CM002, CM003, CM005, CM006, CM008, CM009, CM015]

2.4 Growth drivers, constraints, and diligence gaps

Adoption drivers are visible: manufacturing upgrading, demand for safer human-robot collaboration, expanding embodied-intelligence use cases, and an ecosystem model that tries to reduce integration friction. The latest filing also shows that ROKAE is already converting those drivers into product revenue instead of relying only on custom project work. The main constraints are equally important. Independent commentary describes an intensely competitive domestic market, notes that Chinese vendors have taken majority share in multi-joint robots, and warns that ROKAE is still loss-making while trying to defend leadership in a relatively narrow flexible-collaborative niche. Because the assigned pool lacks buyer-budget documents, procurement examples, and a second independent market-sizing benchmark, the most material diligence gap is not whether the market exists, but whether the serviceable share and payback path can be bounded tightly enough for underwriting.[CM007, CM008, CM009, CM010, CM011, CM012, CM013]

2.5 Exhibits

3.1 The buying set is broader than cobot vendors

A buyer can meet the same automation need through ROKAE, a global cobot platform such as Universal Robots, an established industrial-robot vendor such as FANUC, ABB, KUKA, or Yaskawa, a domestic Chinese supplier, an emerging embodied-AI platform, or an integrator-built fixed cell. The retained evidence does not name or quantify every domestic peer, so the landscape should be read as a category map rather than a league table. ROKAE's breadth across industrial, flexible collaborative, and embodied-intelligent robots is strategically useful because it can address multiple automation jobs within an account. However, Universal Robots' installed base and software ecosystem, and global incumbents' service networks, remain distribution and switching-cost advantages that a capable arm alone does not erase.[CP001, CP005, CP007, CP008, CP009]

3.2 ROKAE is scaling a differentiated hardware stack

ROKAE's current evidence is strongest on its own operating scale. In the six months ended June 30, 2026, the issuer reported RMB150.030 million of industrial-robot revenue, RMB104.910 million of flexible-collaborative-robot revenue, and RMB139.102 million of embodied-intelligent-robot revenue. Robot-product sales volume rose from more than 4,000 units in the comparable 2025 period to more than 13,000 units. The same filing reported 69% year-on-year industrial-robot growth and 72% flexible-cobot growth. Product breadth and integrated joint force control therefore have commercial evidence behind them, not just a prototype narrative. Yet retained sources contain no independent teardown, repeatability benchmark, uptime study, or like-for-like test against named rivals, so capability cells outside ROKAE's own disclosures remain unknown rather than guessed.[CP002, CP003, CP004, CP010]

3.3 Distribution and lifecycle support matter more than public list price

Public pricing is too sparse for a reliable total-cost comparison. A third-party directory estimates Human.X at US$150,000-US$200,000, but explicitly labels that figure a typical class range rather than a ROKAE quotation. No retained source discloses list prices, realized discounts, contract duration, service fees, or integrator economics for ROKAE's industrial and collaborative lines or for named peers. Distribution evidence is more concrete: ROKAE reports parts warehouses in nine countries and service teams across 13 countries and regions. Its registered business scope also spans installation, maintenance, software, and leasing. That breadth can reduce buyer coordination costs, while also making application engineering, controls, software, spares, and service part of the switching decision. Multi-homing and retention rates remain undisclosed.[CP006, CP011, CP012, CP013, CP014]

3.4 The moat is credible in hardware, less proven in ecosystem

ROKAE's strongest moat hypothesis is control over force-sensing joints, motion control, and a product family spanning conventional and emerging robot categories. The reported 34% share of China's flexible-collaborative-robot subsegment is meaningful, but it comes from a Frost & Sullivan study commissioned for the listing and the underlying methodology is not in the retained evidence. The broader market is becoming more competitive as domestic manufacturers gain share against foreign incumbents. ROKAE also reported an RMB79.0 million HKFRS loss for the first half of 2026 despite an RMB18.4 million adjusted non-HKFRS profit, preserving the distinction between accounting performance and management's adjusted view. International expansion adds certification, channel, training, documentation, and local service costs. The investment case therefore depends on converting hardware differentiation and volume growth into repeat deployments before competitors commoditize the arm and shift value toward software and integration ecosystems.[CP007, CP008, CP010, CP015, CP016]

3.5 Exhibits

4.1 Revenue model and public traction

The clearest reader-facing financial improvement is mix: ROKAE reported H1 2026 revenue of RMB414.7 million, with sales of robotic products contributing RMB394.0 million, or 95.0% of the total, while project-style robotic solutions fell to RMB9.0 million after management said it became more selective on new projects. Within product revenue, embodied intelligent robots already contributed RMB139.1 million, or 33.5% of total revenue, which matters because the strongest margin path appears to come from scaling standardized robot products rather than bespoke solutions work. Public traction is real but incomplete: official sources cite more than 1,000 customers, more than 40 countries, and 50,000 robots of annual production capacity, yet they do not expose the realized ASP, repeat-purchase cadence, or backlog coverage needed to judge how sticky that activity is.[CI001, CI002, CI008]

4.2 Margin path and unit-economics visibility

The H1 2026 filing and the later earnings summary both point to real gross-margin improvement: gross profit rose to RMB123.4 million, reported gross margin reached 29.8%, and product gross margin before inventory write-downs reached 31.1% as higher unit volumes reduced unit costs and improved supplier bargaining power. That is a meaningful signal that the company is moving away from lower-margin solution delivery toward more repeatable product economics. The caution is that almost every underwriting variable below gross margin remains opaque in public: retained sources do not disclose realized robot pricing, discounting, CAC, payback, funnel conversion, sales-cycle length, churn, or backlog, so the public record is good enough to show improving economics but not good enough to model mature sales efficiency.[CI003, CI009, CI012]

4.3 Liquidity, balance sheet, and financing dependency

ROKAE's capital position improved materially after the reporting period, but the timing matters. As of June 30, 2026, the company reported RMB115.9 million of cash and cash equivalents, RMB138.9 million of term deposits, and RMB132.9 million of financial assets at fair value through profit or loss, while shareholders' deficit still stood at RMB1,221.7 million and borrowings were RMB50.6 million. The July-August 2026 Hong Kong IPO and greenshoe then added approximately HK$937.1 million of net proceeds earmarked through FY2029 for R&D, overseas sales build-out, production capacity, strategic investment, and working capital. That means near-term liquidity pressure is meaningfully lower than the June balance sheet alone suggests, but capital adequacy still depends on management converting fresh equity into sustained margin expansion rather than simply funding a longer loss runway.[CI004, CI005, CI006, CI007, CI011]

4.4 Financial verdict and diligence blockers

The financial verdict is directionally positive but not yet underwriteable on public evidence alone. Revenue quality improved because product sales dominated H1 2026 mix, gross margin expanded, and operating cash flow turned positive, while post-period IPO proceeds reduce immediate financing risk. Even so, independent reporting still frames ROKAE as loss-making, internationally underpenetrated, and reliant on continued R&D spending to defend share in a still-small collaborative-robot niche. Before treating the recent step-up as a durable margin inflection, diligence should request model-level ASPs, discounting, backlog, exact cash-flow bridge, customer-concentration updates, and channel/sales-efficiency data that are absent from the retained public corpus.[CI010, CI012, CI013]

4.5 Exhibits

5.1 Product scope and workflow fit

The retained evidence supports treating ROKAE as a multi-line robotics platform rather than a single-model cobot vendor. Official company and investor surfaces consistently group the offering into industrial robots, flexible collaborative robots, embodied intelligent robots, and related robotic solutions, while the investor-relations profile adds that the public lineup already spans 50+ robot products. The workflow focus is also specific rather than generic: collaborative robots are marketed for high-precision assembly, compliant human-machine interaction, welding, and medical or rehabilitation-adjacent tasks, while the embodied portfolio is aimed at humanoid-arm, wheeled dual-arm, and related physical-AI development contexts. That breadth matters for diligence because it suggests ROKAE is trying to monetize not only robot bodies but also deployment-ready workflows across industrial, commercial-service, medical, and embodied-AI scenarios.[CE001, CE002, CE003, CE012]

5.2 Architecture and integration surface

The public architecture story has three visible layers. At the hardware layer, ROKAE markets industrial robots, xMate collaborative robots, AR humanoid force-controlled arms, and HSA force-controlled joints. At the control layer, 2026 disclosures describe xCore and iNex as the motion-and-interface substrate and position ROKAE Brain plus force-perception algorithms as the embodied-intelligence "brain" layer. At the external integration layer, the strongest non-marketing proof is the ROS2 manual: ROKAE documents supported xMate models, ROS2 Humble on Ubuntu 22.04, MoveIt 2 plus ros2_control integration, and a hardware plugin that calls a vendor SDK. That makes the platform more concrete for integrators than a simple brochure would, but it also shows that third-party deployment still depends on ROKAE-controlled APIs, model support, and documentation quality rather than a fully open ecosystem.[CE004, CE006, CE007, CE008, CE009, CE010, CE011]

5.3 Maturity, controls, and product risk

ROKAE's public trust story is strongest on functional safety and deployment convenience. The company repeatedly cites TÜV-backed functional-safety features, ISO collaborative-robot standards, redundancy in sensing and control, simulation tooling, and a global service network. Independent reporting also repeats the self-developed control-stack and force-control differentiation themes, which gives some external corroboration that these are the right things to diligence. Still, the public evidence is incomplete where an investor would want it to be firmest: the filing says ROKAE still plans to obtain mainstream certifications such as CE and UL and to strengthen localized service capabilities, while the retained pool contains no direct evidence on controller security architecture, software update governance, privacy handling, or fleet-level uptime/warranty outcomes by model. The result is a chapter that can verify product breadth and architecture direction, but not yet underwrite global deployment quality with high confidence.[CE005, CE006, CE011, CE013]

5.4 Exhibits

6.1 Customer footprint is broad, but still manufacturing-heavy

The assigned pool supports a broad but still mostly manufacturing-centered customer picture. Company surfaces consistently frame ROKAE around industrial robots, collaborative robots, and embodied-intelligence products sold into auto parts, electronics, metals, healthcare, commercial service, and adjacent factory workflows, while the interim filing says the installed base has reached more than 1,000 enterprises across more than 40 countries and regions. The same filing shows the commercial footprint is still domestically concentrated in revenue terms: Mainland China accounted for RMB387.6 million of first-half 2026 customer revenue versus RMB27.1 million overseas. That mix matters because the customer story is stronger on deployment breadth than on balanced geographic monetization. [CU001, CU002, CU003]

6.2 Proof of use is strongest in automotive and high-end manufacturing

The clearest production-grade proof in the pool comes from IDEC Factory Solutions' inspection report on ROKAE installations in China, which describes about 30 ROKAE robots operating at a Cummins factory across assembly, deburring, spot welding, and inspection, and reports a two-year payback. Official 2026 event coverage broadens that proof set rather than fully replacing it: ROKAE showcased more than ten application solutions at ITES Shenzhen across 3C electronics, automotive, and precision machining, while Auto China coverage described an embodied-robotics turnkey stack positioned to move partners from technical validation toward commercial deployment. Together, those sources support real adoption surfaces, but they do not yet provide a deep roster of named recurring customers by revenue band or contract duration. [CU004, CU005, CU006, CU007, CU008, CU009]

6.3 Expansion signals exist, but retention and concentration proof is still thin

Expansion evidence exists in three forms. First, the interim filing says certain existing customers repurchased relatively mature solutions in the first half of 2026. Second, the same filing says a number of leading embodied-intelligence technology enterprises adopted or switched to ROKAE humanoid robotic arms for R&D and product iteration. Third, the broader capital-markets narrative includes explicit plans to expand the overseas sales network. Even so, the assigned pool does not disclose NRR, GRR, churn, renewal rates, average contract length, or top-customer revenue share. The result is a customer chapter with decent deployment and channel proof, but incomplete durability and concentration evidence for underwriting recurring revenue quality. [CU010, CU011, CU012, CU013, CU014]

6.4 Exhibits

7.1 Ranked risk posture

The most material near-term risk is still financial durability rather than an immediate legal overhang. H1 2026 revenue, adjusted profit, and operating cash flow improved sharply, but ROKAE still reported a RMB79.0 million statutory loss, while independent commentary says losses may continue and that gross-margin expansion still has to prove a durable path to profitability. Regulatory and expansion risk sits close behind: management itself says it is monitoring cross-border trade regulations and geopolitical dynamics while building overseas service capacity, and independent analysis flags exposure to export controls and sanctions as the company internationalizes. The visible mitigations are real but incomplete: post-IPO proceeds are earmarked for R&D, market expansion, working capital, and production efficiency, yet public evidence still leaves material diligence gaps around certifications, channel redundancy, and process controls.[CR003, CR004, CR005, CR006, CR010, CR013, CR014]

7.2 Legal and regulatory exposure is visible, but not fully diligenceable

The cleanest current legal fact is narrow: as of June 30, 2026, the interim results announcement says ROKAE had no material contingent liability and no pending or threatened litigation or claim of material importance. That helps on near-term overhang, but it does not prove broad compliance readiness. The formal legal framework is visible because the articles of association explicitly anchor the company to PRC company and securities law plus Hong Kong listing rules, while independent IPO analysis adds a concentrated post-listing control structure around founder-chairman-CEO Tuo Hua. The harder question is operational compliance as the company expands abroad: management highlights cross-border trade and geopolitical monitoring, and independent analysis explicitly names export controls and sanctions, yet the retained public corpus still does not enumerate the underlying compliance, certification, privacy, or safety-control packet needed to judge maturity.[CR001, CR002, CR003, CR004, CR013, CR015]

7.3 Operating execution depends on converting product breadth into repeatable delivery

Public evidence shows real operating scale, but it also shows how much execution still depends on R&D translation and field delivery. Official materials point to a 50,000-robot annual factory capacity, service coverage in mainland China, overseas parts warehouses, and technical support teams across 13 countries and regions, so this is not a paper platform. At the same time, the company is still investing aggressively in control systems, force-controlled joints, perception, and embodied AI commercialization, while independent sources warn that maintaining share in a still-small collaborative-robot segment will require sustained R&D intensity and that planned overseas service expansion has not yet proven traction. Open SDK, ROS2, plugin, and control-system documentation are a credible mitigation because they lower integration friction, but they also imply a support, reliability, and ecosystem-adoption burden that public sources do not quantify.[CR004, CR006, CR007, CR011, CR012, CR013]

7.4 Balance-sheet relief after listing does not remove financing and counterparty dependence

The July 2026 listing reduced one important legacy overhang because preferred-share redemption liabilities were derecognized into equity after the IPO, but the public evidence still describes a business that relies on careful counterparty management. The filing says credit risk mainly arises from customer credit extension and investing activities, and it also shows RMB50.648 million of bank loans under supplier-finance arrangements that settle within one year. Those exposures matter because profitability is not yet fully proven and because the post-IPO use-of- proceeds plan still reserves meaningful capital for working capital, production efficiency, and strategic investment rather than only discretionary expansion. The practical question for diligence is whether improved H1 2026 cash generation is a durable operating shift or a temporarily favorable mix before global scale, service redundancy, and collections discipline are fully tested.[CR005, CR008, CR009, CR010, CR014]

7.5 Exhibits

8.1 Price support exists, but proof still trails the valuation

The current valuation context is tangible rather than hypothetical. Independent IPO analysis put the offer at HK\$38.00 per H share for 23,031,900 new shares and an implied market capitalization of roughly HK\$9.95 billion, while post-debut reporting said the July 2026 listing raised HK$875.2 million in a heavily oversubscribed transaction. That level of demand is easier to respect because the operating print also improved: H1 2026 revenue reached RMB414.7 million, gross margin rose to 29.8%, adjusted net profit turned positive, and operating cash flow turned positive during the period. Even so, valuation support is incomplete rather than complete. Public sources still show a RMB79.0 million statutory loss, and the retained pool does not provide an FX-normalized peer multiple bridge, a priced private valuation mark, or enough post-listing trading history to defend paying up simply because the book was strong.[CV001, CV002, CV003, CV004, CV014, CV016]

8.2 Financing risk eased, but governance and cap-table visibility remain partial

Near-term financing pressure is lower than it was pre-listing. Retained sources show that the company directed roughly 40% of IPO net proceeds to R&D and another 30% to overseas sales-network expansion, while lock-ups kept existing shareholders, ESOP platforms, and pre-IPO investors restricted for 12 months after listing. The same retained IPO analysis says special rights granted to pre-IPO investors were terminated on listing, which reduces the most obvious preference overhang. The remaining caution is governance concentration and incomplete historical funding visibility. Independent analysis still describes founder Tuo Hua as controlling approximately 38.78% of aggregate voting power, and retained sources do not reconcile the full pre-IPO financing bridge cleanly: one independent analysis cites more than RMB1.4 billion across ten rounds, while a startup-profile page repeats superlative positioning without publishing the priced marks or cap-table terms that would let investors judge dilution discipline directly.[CV007, CV008, CV009, CV010, CV011, CV012, CV015]

8.3 Scenario framing is possible, but comparable discipline is still weak

A scenario framework can be grounded in retained evidence, but a precise valuation range cannot. The bull case rests on order and shipment momentum converting into repeatable commercialization of embodied-intelligence products, with enough IPO-funded reinvestment to keep growth and margin moving together. The bear case is equally visible: retained independent evidence still flags ongoing losses, declining R&D intensity as a share of revenue, technology and customer concentration risk, and legal or export-control exposure. Comparable evidence is the weakest part of the public record. Independent listing coverage places ROKAE inside a broader Chinese robotics listing wave alongside Unitree, UBTech, and X-Variable, but the retained sources do not provide normalized revenue, margin, or cash-flow baselines for those names, so they are milestone references rather than clean valuation comps.[CV005, CV007, CV012, CV013, CV014, CV015]

8.4 Recommendation is track / research-more until price support deepens

The evidence-weighted recommendation is not "avoid the company" but "avoid underwriting the current valuation as proven." ROKAE already looks like a scaled platform business by Chinese robotics standards: it serves over 1,000 customers in more than 40 countries and regions, runs a 50,000-robot annual-capacity factory, and has enough recent demand evidence to make the story investable in principle. What the public record still cannot do is close the gap between operational momentum and entry discipline. Without a normalized comparable set, a priced private valuation history, clearer cap-table economics, and harder evidence on ASPs, backlog conversion, and post-listing profitability, the right call is track/research-more with a high risk rating and a price-sensitive stance. A cleaner buy case would require the company to grow into the valuation through additional periods of positive operating cash flow and narrower statutory losses rather than through narrative alone.[CV003, CV004, CV006, CV014, CV015, CV016]

8.5 Exhibits