Linkerbot

1.1 Identity and business model

Linkerbot is a Beijing-based robotics company founded in 2023 and focused on dexterous robotic hands, associated control software, and manipulation-skill infrastructure. The narrow component strategy lets it sell into research, humanoid-robot, and industrial workflows without underwriting the timing of a complete general-purpose humanoid. Preqin describes direct sales of robotic products to enterprises and institutions as the primary revenue model, while independent coverage describes a portfolio spanning hand hardware and skills software. That combination suggests a hardware-led business with a possible data and software layer, but no source in the assigned evidence discloses revenue mix, recurring software economics, gross margin, or customer concentration.[CO001, CO002, CO015]

1.2 Leadership and governance

Founder evidence is directionally consistent but not fully normalized. RobotsBeat identifies Alex Zhou as the founder, while KrASIA identifies Zhou Yong as founder and CTO and describes his earlier Lenovo research experience. The two names may refer to the same person, but the assigned sources do not establish that identity explicitly. KrASIA also names Cao Gang as the leader for algorithms and model training, which adds technical depth beyond the founder. No retained source provides a board roster, voting rights, succession plan, executive ownership, or leadership-change history. The investment case therefore has visible founder and technical leadership but material governance and key-person diligence gaps.[CO003, CO004, CO005]

1.3 Financing and stakeholders

The financing record shows unusually rapid capital formation but does not yet reconcile into a dependable total-raised or dilution history. Late-2025 reporting described a nine-figure-renminbi Series A+ and a fourth financing event in eight months. Tech in Asia then reported a RMB 1.5 billion Series B in February 2026, and Preqin recorded a further Series B+ financing in April. Coverage places the Series B+ valuation at $3 billion; a later $6 billion figure is a target for a proposed new round, not evidence of a completed transaction. Financial and strategic investors are numerous, but ownership, preferences, board rights, secondary sales, and debt remain undisclosed.[CO006, CO007, CO008, CO009, CO010]

1.4 Scale, operating signals, and milestones

Operating signals are substantial but mostly issuer-originated or based on management statements. Linkerbot said it shipped 10,000 hands in 2025 and represented 80% of worldwide demand; KrASIA separately reported monthly orders above 1,000 units, four subsidiary factories, and monthly capacity above 4,000 units. The company also set a 2026 delivery ambition of 50,000 to 100,000 hands. Product breadth spans three actuation architectures and 11 to 42 degrees of freedom, supporting several price and performance tiers. These milestones indicate commercialization momentum, but they do not substitute for audited shipments, accepted-customer deliveries, revenue, utilization, yield, or warranty data. A potential Hong Kong IPO adds a near-term disclosure catalyst, although the reported process remained changeable.[CO011, CO012, CO013, CO014, CO015, CO016]

1.5 Exhibits

2.1 A component market, not the whole humanoid economy

The investable boundary is high-degree-of-freedom robotic hands and the closely coupled controls, teleoperation, and skills software needed to deploy them. Linkerbot's own catalog spans linkage, direct-drive, and tendon-driven hands with materially different degrees of freedom, weights, loads, and interfaces, then extends into arms, a humanoid platform, motion capture, and force feedback. That breadth makes a broad "robotics" denominator misleading. The core market should include hand hardware and hand-specific integration; it should exclude complete humanoid bodies, general-purpose robot arms, unrelated factory automation, and model-compute spend. Conventional grippers, internal end-effector development, and waiting for a more mature humanoid stack remain economically relevant substitutes rather than addressable dexterous-hand revenue.[CM001, CM002, CM003, CM014]

2.2 Commercial proxies are useful, but they are not TAM

Available evidence offers shipment, order, capacity, pricing, and claimed-share signals rather than a defensible market-size series. Linkerbot's CES release claims more than 80% of the global high-DOF hand market and thousands of units of monthly production. China Biz Insider separately reported monthly orders above 1,000 units, monthly capacity above 4,000, and an aggressive 2026 delivery target, but those figures trace back to company statements and do not disclose the category denominator. Published prices span subsidized entry models through research-oriented products, demonstrating segmentation while also showing why one unit count cannot be multiplied by one list price. A numeric TAM, SAM, or SOM would therefore create false precision until independent shipment data, realized average selling prices, and a stable definition of "high-DOF" are obtained.[CM004, CM005, CM008, CM009, CM010]

2.3 Buyer, user, and payer vary by deployment

Research laboratories buy controllable hardware and data-collection tooling; humanoid OEMs buy a subsystem that must fit their arm, control stack, and target tasks; industrial users ultimately pay for completed work rather than degrees of freedom. Reporting places Linkerbot with customers in scientific research, humanoid robotics, and industry, and company-issued materials name Samsung, Siemens, and Foxconn. Those references show segment breadth but not procurement depth, repeat orders, or customer economics. Near-term adoption can occur on existing robotic arms before general-purpose humanoids mature, which expands the immediate buyer set. Budget ownership should nevertheless be validated deployment by deployment: robotics R&D may fund pilots, while operations or automation teams must justify scaled use through integration effort, task success, service life, and replacement cost.[CM006, CM007, CM012]

2.4 Falling entry prices broaden trials; reliability governs scale

The principal growth drivers are more affordable hands, multiple architectures for different payload and control requirements, and software that converts demonstrations into reusable manipulation skills. Linkerbot says LinkerSkillNet contains more than 500 real-world skills, and its product range gives buyers options from lower-cost entry hardware to high-DOF research systems. The binding constraints are less visible: integration into third-party arms, repeatable task completion under variable conditions, maintenance, replacement cost, safety, and proof that demonstrations survive production workloads. A large technology company could also use scale to compress hardware margins. Financing and a possible IPO may fund capacity, but capital availability does not independently validate end-market size. Valuation work should therefore weight verified deployments and unit economics above claimed category share.[CM003, CM011, CM012, CM013]

2.5 Exhibits

3.1 Competitive landscape and buyer alternatives

Linkerbot competes in a layered market rather than against one clean peer set. Shadow Robot is the legacy high-articulation benchmark: Pandaily reports 24 independently controlled joints and a roughly $140,000-$280,000 price range. Tesla, Boston Dynamics, Figure AI, 1X, and Apptronik represent a different threat because they can internalize hand development inside a complete humanoid platform. At the lower-complexity end, conventional commercial grippers with one to six degrees of freedom remain adequate for constrained pick-and-place work. Buyers can also retain current automation or build an end effector internally. Linkerbot's component-supplier model therefore expands its addressable customer set beyond any single robot body, but also leaves distribution and integration power with robot OEMs and automation integrators.[CP001, CP003, CP004, CP013]

3.2 Capability breadth and product position

Linkerbot's most defensible product distinction is architectural breadth, not degree-of-freedom count alone. Multiple reviewed sources describe linkage, direct-drive, and tendon-driven products, allowing the company to trade among stiffness, compactness, speed, maintainability, and biomimicry. The O6 also establishes a low advertised entry point: a September 2025 report listed it at RMB 6,666, while Pandaily described the L20 at roughly one-twentieth the cost of comparable overseas products. Those figures are useful positioning signals, not proof of delivered cost per successful task. RobotsBeat correctly cautions that active and passive joint counts do not establish reliability or maintenance burden, and Linkerbot's own co-founder acknowledged that earlier emphasis on peak metrics produced inconsistent stability and difficult customer deployments.[CP002, CP005, CP006, CP011, CP012]

3.3 Pricing, distribution, and switching economics

Public evidence supports list-price and architecture comparisons, but not realized transaction pricing, discounts, integration labor, service life, or replacement cost. Linkerbot's software and data layer may create switching cost if customers standardize workflows around LinkerSkillNet: company-originated material says the library has more than 500 skills, while RobotsBeat describes teleoperation, a skills store, and a model layer around the hardware. The same coupling can work against adoption when buyers already own a control stack or prefer an integrated robot OEM. Capital access may help Linkerbot subsidize manufacturing and channel expansion; Tech in Asia reported a RMB 1.5 billion Series B and a possible Hong Kong IPO, while Preqin records a subsequent Series B+ round but does not disclose total capital in its public preview.[CP007, CP014, CP015]

3.4 Moat durability and displacement risk

The emerging moat thesis combines cost-down manufacturing, multi-architecture engineering, installed hardware, and proprietary manipulation data. Company-originated releases claim thousands of units per month, more than 500 skills, and over 80% share of high-degree-of-freedom hands; Gasgoo separately reports a 60,000-100,000-unit annual-capacity line and more than 140 CNC machines, but its underlying operating metrics appear supplied by Linkerbot rather than audited. These signals support a scale hypothesis, not a verified market-share conclusion. Displacement risk remains meaningful from full-stack humanoid makers, well-capitalized component entrants, and simpler grippers that solve narrower jobs more reliably. The investment case should therefore treat verified task success, service life, gross margin, customer retention, and independently measured share as gating evidence rather than extrapolate a durable moat from shipment and demonstration claims.[CP008, CP009, CP010, CP016]

3.5 Exhibits

4.1 Revenue architecture and public pricing

The clearest public evidence describes a hardware-led business: Preqin says Linkerbot generates revenue primarily through direct sales of robotic products to enterprises and institutions, while independent reporting describes customers mounting its hands on existing robotic arms rather than purchasing complete humanoids. This broadens the near-term buyer base beyond full humanoid manufacturers, but it still leaves contract structure, acceptance terms, warranty obligations, and revenue recognition undisclosed. An independent product review reports starting prices from about $1,500 for the O6 through about $14,000 for the L30, with intermediate L-series tiers. Those are public starting prices, not evidence of realized average selling price, discounts, revenue mix, or gross margin. LinkerSkillNet may strengthen the hardware proposition, but no retained source establishes a separate software subscription, license, or usage-fee stream.[CI001, CI002, CI003, CI004, CI005, CI006, CI007, CI008]

4.2 Production traction is visible; unit economics are not

Reuters reported production of almost 5,000 hands per month and a plan to reach 10,000 soon, a scale signal that is commercially meaningful but is neither revenue nor proof of utilization, sell-through, or profitable output. Linkerbot's own CES release claimed a two-week delivery cycle that was six times faster than competitors, yet the release supplies no audited baseline, order backlog, cancellation rate, or on-time-delivery evidence. The manufacturing footprint is also not fully reconciled: Reuters reported five factories and more than 400 employees, whereas the company release described four proprietary factories. Together these disclosures indicate a meaningful physical cost base involving factories, inventory, labor, quality control, warranty service, and likely working capital. They do not reveal gross margin, contribution margin, capex, inventory turns, warranty reserves, CAC, payback, or cash conversion.[CI009, CI010, CI011, CI012]

4.3 Funding signals do not establish capital adequacy

Public financing reports show access to capital, but they cannot be treated as cash on hand. One independent review reported $150 million raised since April 2025, while Tech in Asia reported a February 2026 Series B of CNY1.5 billion, or about $220 million. These disclosures are not reconciled and may overlap. Reuters then reported that a Series B+ valued Linkerbot at $3 billion and that management was targeting $6 billion for a subsequent financing; independent coverage correctly distinguishes the target from a completed transaction. Tech in Asia and The Robotics Media also reported a possible Hong Kong IPO that could raise a few hundred million dollars, but both characterize the process as prospective. None of these figures discloses current liquidity, monthly burn, committed capex, debt, project finance, restricted cash, or the milestone that would trigger the next raise.[CI013, CI014, CI015, CI016]

4.4 Financial verdict and diligence threshold

The public record supports a plausible direct hardware-sales model, unusually high claimed production scale, and continued financing access. It does not support a revenue-quality or solvency conclusion. Starting prices may indicate attractive revenue per unit, but there is no realized-price distribution, paid shipment count, customer concentration, returns data, service burden, or margin bridge. Similarly, factory expansion may create operating leverage at volume or may consume cash through equipment, inventory, and support; the available evidence cannot choose between those outcomes. Underwriting should therefore remain conditional on bank statements and a monthly financial package reconciling orders, accepted shipments, recognized revenue, bill of materials, labor, warranty, gross margin, working capital, capex, cash, debt, and burn. The next financing or IPO should be modeled as a dependency, not as available capital, until proceeds and closing conditions are documented.[CI002, CI003, CI004, CI005, CI006, CI009, CI010, CI011, CI012, CI013, CI014, CI015, CI016]

4.5 Exhibits

5.1 Product portfolio and customer job

Linkerbot's core product is a family of dexterous robotic hands used as end-effectors on humanoid and industrial robots, rather than a single standardized hand. The range uses linkage, direct-drive, and tendon-driven mechanisms, allowing buyers to trade active control, passive compliance, weight, payload, precision, and cost. Public listings describe a wider ecosystem that includes a collaborative arm, a humanoid platform, motion-capture and force-feedback gloves, a skills library, and model software. This creates a credible customer workflow from selecting a hand through collecting demonstrations and teaching manipulation skills. However, model names and specifications vary across third-party listings, so procurement should be tied to current datasheets, configuration-specific acceptance tests, and a bill of materials rather than family-level marketing.[CE001, CE002, CE003, CE004, CE005, CE006, CE007, CE008]

5.2 Architecture from actuation to learned skills

The available architecture is layered. At the physical layer, product families use different transmission mechanisms and combine active and passive joints. Sensing claims include force-torque measurement, tactile arrays, cameras, and electronic skin. Linkerbot also describes an integrated electric-cylinder drive and vertically controlled manufacturing inputs. Above the hardware, Open TeleDex is presented as a modular teleoperation system, while LinkerSkillNet supplies a skills library and Dex-Serl supports real-machine reinforcement learning. This is strategically coherent because the hand can generate data as well as execute learned behavior. It is not yet a verified closed-loop architecture: the retained evidence does not disclose APIs, timing budgets, model evaluation, deployment topology, update controls, or interoperability test results.[CE002, CE007, CE009, CE010, CE011, CE012, CE015, CE016]

5.3 Deployment, integration, and use cases

The product can enter customer workflows at several points. Research teams can use a hand and gloves for data collection; robot makers can integrate a hand with an existing arm; manufacturers can evaluate the L20-class linkage design for tool use and variable-object handling; and demonstration workstations can combine two arms with dexterous hands. RobotToday names universities and industrial companies associated with the products, while Tech in Asia reports that management is targeting nearer-term demand from existing robotic arms rather than waiting for mature humanoid fleets. These signals show plausible deployment surfaces, not proven production economics. Public sources do not provide installation time, supported controllers, environmental ratings, spare-parts policy, field-service coverage, task success rates, or customer-owned before-and-after outcomes.[CE001, CE003, CE007, CE009, CE010, CE013, CE014]

5.4 Maturity, differentiation, and technical risk

Linkerbot's strongest visible differentiation is breadth: several actuation architectures, lightweight and high-DoF variants, integrated sensing, and a data-and-skills layer. Company-issued material further claims four factories, micrometer-level tolerance control, a two-week delivery cycle, and thousand-unit monthly production. Those assertions indicate an industrial ambition but remain supplier claims rather than independent process audits. The retained evidence contains no published MTBF, duty-cycle failure curves, ingress or temperature qualification, maintainability data, cybersecurity architecture, privacy controls for teleoperation data, functional-safety certification, recall history, or third-party comparative benchmark. Consequently, the investment case should treat scale, durability, safety, and data defensibility as testable diligence hypotheses rather than verified moats.[CE004, CE006, CE008, CE009, CE012, CE013, CE015, CE016]

5.5 Exhibits

6.1 Buyer and user segments

The clearest supported customer split is between research users and industrial buyers. China Robotics Insider describes manipulation and teleoperation research as an application, while RobotsBeat says research teams need learning hardware and manufacturers need end effectors that fit existing arms and processes. Tech in Asia further reports that Linkerbot is pursuing nearer-term demand by selling hands for installed robotic arms rather than waiting for mature humanoid platforms. This creates a component-led route to market across laboratories, automation teams, and robot integrators, but the reviewed evidence does not quantify revenue, units, geography, or buyer mix by segment.[CU001, CU002, CU011]

6.2 Adoption signals and named-account proof

Public scale signals are substantial but heterogeneous. A December 2025 report relaying company disclosures says monthly orders exceeded 1,000 units and gives a 2026 delivery target of 50,000 to 100,000 hands. The Next Web, citing Reuters, separately describes the company as capable of shipping more than 1,000 units per month. A Linkerbot release names Samsung, Siemens, and Foxconn as customers, and Tech in Asia repeats Samsung and Siemens. These data points indicate demand and recognizable counterparties, but they do not provide a common period, shipped-versus-ordered reconciliation, contract values, or customer-side confirmation of production deployment.[CU003, CU004, CU005, CU006, CU007, CU014, CU015]

6.3 Durability, reliability, and concentration

The strongest adverse evidence comes from Linkerbot's own account in a FedEx Business Insights interview. The company says an earlier focus on peak performance produced inconsistent stability, difficulty sustaining continuous operation, and harder-than-expected customer deployment; it says priorities then shifted toward continuous operation, thermal stability, maintenance, and practical task-completion metrics. This is relevant evidence of learning, not proof that the problems are resolved. None of the eight retained sources supplies NRR, GRR, churn, renewals, contract duration, customer cohorts, top-customer share, or channel concentration, leaving durability and dependency risk unquantified.[CU008, CU009, CU012, CU013]

6.4 Expansion path and investment implication

Linkerbot's plausible expansion loop begins with a hand selected for research or an existing robotic arm, moves through integration and task qualification, and then depends on reliable production use before additional hardware, skills, or sites can follow. The reported RMB 6,666 O6 price could reduce an entry barrier, but it is a September 2025 reported price with no configuration, service, integration, or lifecycle-cost denominator. Linkerbot also says its hands assemble robotic arms and other hands internally; that is an operational proof point rather than external customer validation. Investment underwriting should therefore require customer-side references and cohort data before treating shipments as durable land-and-expand revenue.[CU010, CU011, CU016]

6.5 Exhibits

7.1 Ranked risk posture

The highest residual exposure is operational reliability at scale. In a July 2026 interview, Linkerbot's co-founder acknowledged that earlier systems showed inconsistent stability, difficulty sustaining continuous operation, and harder-than-expected customer deployment; management says it has since shifted attention toward continuous operation, thermal stability, maintenance, and practical task completion. That response is directionally appropriate, but the assigned evidence contains no independently measured field-failure, uptime, warranty, or return-rate series. A second high exposure is scale execution: a company-reported 10,000th cumulative shipment in December 2025 sits against a reported target of 50,000-100,000 deliveries in 2026. The risk is not that the target is impossible, but that quality controls, service capacity, and working capital are not publicly evidenced at the same level as the volume ambition.[CR001, CR002, CR005, CR007, CR008]

7.2 Operational, quality, and safety exposure

Product demonstrations and degree-of-freedom specifications do not establish dependable customer outcomes. RobotsBeat explicitly notes that active and passive joint counts alone cannot prove task reliability or maintenance burden, while Linkerbot's own humanoid page describes a roughly 100 kg mobile platform, a 2 kg arm-end payload, and remote operation from a safe distance. Those disclosures make controls such as obstacle sensing and teleoperation visible, but they do not substitute for third-party functional-safety certification, incident history, or autonomous task-completion evidence. The O6's reported 6,666 RMB price may widen adoption, yet without gross-margin, warranty-cost, service-life, and replacement-cost data it also creates an unresolved economics and quality tradeoff. Investment underwriting should therefore require cohort-based reliability and cost data, not peak demonstrations.[CR001, CR002, CR003, CR004, CR012, CR013, CR015]

7.3 Regulatory, legal, and governance diligence

The assigned public evidence does not establish Linkerbot's certification coverage, product-liability controls, export-control posture, privacy and training-data rights, patent disputes, recalls, or litigation status. That absence is a diligence gap, not evidence that no issue exists. BestStartup.Asia separately flags geopolitical scrutiny arising from state-fund investors and a Beijing-Shenzhen footprint, while Tech in Asia reports that a possible Hong Kong IPO remained under discussion and could change. An IPO process may improve disclosure discipline, but it does not currently provide audited operating or risk data in this source set. The regulatory register is therefore deliberately partial and uses conditional severity rather than invented incident probabilities.[CR003, CR004, CR009, CR010, CR011]

7.4 Dependencies, financing, and thesis-break triggers

Linkerbot depends on robot integrators continuing to buy a specialized hand rather than vertically integrating it. BestStartup.Asia identifies Tesla's in-house hand work and whole-robot ambitions among platform companies as direct strategic threats; it also frames the component supplier's ceiling as sensitive to whether humanoids remain a niche. The company's claimed 80% share is therefore concentration in a young category, not automatic durability. Financing evidence also needs reconciliation: a December 2025 release cited $150 million raised since April 2025, Tech in Asia later reported a $220 million February 2026 Series B, and other reporting described a proposed $6 billion valuation after a $3 billion Series B+. Kill criteria should include unresolved audited funding data, a material rise in field failures, failed production qualification, or major integrators shifting decisively in-house.[CR006, CR009, CR010, CR011, CR014, CR016]

7.5 Exhibits

8.1 Recommendation and entry discipline

The recommendation is research-more, not buy. Independent reporting places a completed Series B+ financing at roughly US$3 billion while describing a US$6 billion next-round or IPO ambition as prospective; Tech in Asia also says advisers were hired for a possible Hong Kong offering whose plans could still change. Those are useful liquidity signals, but they do not establish a defendable entry price. The reviewed evidence lacks audited revenue, margin, cash, burn, ownership, liquidation-preference, and dilution data. Entry discipline should therefore be milestone-based: require a cap table, financing documents, audited financials, customer-level shipment and payment evidence, and a reconciled bridge from operating performance to the proposed valuation before underwriting a target return.[CV001, CV004, CV005, CV006, CV014, CV015]

8.2 Thesis and anti-thesis

The thesis is that a focused hand supplier can monetize before general-purpose humanoids mature. Company-issued releases claim 10,000 cumulative hand shipments, more than 80% share of the high-degree-of-freedom segment, production above 1,000 units per month, and customers including Samsung, Siemens, and Foxconn. Preqin describes direct product sales to enterprises and institutions, while an independent review cites a 6,666 yuan O6 price point. The anti-thesis is that these commercial signals are mostly company assertions or secondary summaries rather than audited evidence. RobotsBeat specifically cautions that specification counts do not establish task reliability or maintenance burden, and that a fundraising ambition is not a completed valuation event.[CV007, CV008, CV009, CV010, CV011, CV012, CV013]

8.3 Bull, base, and bear framework

Public evidence is insufficient for a numerical valuation range, so the scenarios use observable underwriting gates rather than fabricated multiples. A bull case requires independent confirmation of shipment volumes, repeat customers, usable gross margins, and a financing or listing near the reported target without punitive preferences. The base case keeps commercial traction as plausible but unverified and treats the last reported US$3 billion mark as context, not fair value. The bear case is triggered by failed shipment reconciliation, weak paid deployment, costly reliability support, a down round, or an IPO delay. No scenario receives a numerical probability because the evidence set lacks the operating and capitalization inputs needed to calibrate one.[CV004, CV005, CV006, CV007, CV010, CV013, CV015]

8.4 Exit readiness and final diligence asks

Hiring China Merchants Bank International, Citic Securities, and HSBC for a possible Hong Kong offering indicates preparation, not readiness or completion. A financing-led exit could improve price discovery, but current reporting says talks remain ongoing and plans may change. Final diligence should obtain the latest capitalization table, all preference and debt terms, audited financial statements, monthly bookings-to-cash reconciliation, customer concentration and returns data, warranty and field-failure cohorts, and written adviser timelines for listing readiness. Thesis-break triggers are a valuation step-up without corresponding audited economics, material divergence between orders and cash collections, or inability to validate the shipment and customer claims that anchor the growth narrative.[CV001, CV005, CV006, CV007, CV010, CV013, CV014]

8.5 Exhibits