Tianji

1.1 Identity and operating scope

Public company materials and startup databases agree that Tianji is a Dongguan / Songshan Lake robotics company formed in 2015, but they describe the business through the lens of its current force-control niche rather than through audited financial scale. The official site positions Tianji as a robot and control-systems vendor spanning collaborative, articulated, SCARA, and force-controlled humanoid-arm products on its Tianji Fusion control system. That matters for the rest of the report: Tianji should be treated as an automation supplier evolving into embodied-intelligence infrastructure, not as a pure software company or a generic humanoid brand. Scale signals are directionally positive but dated. The official site still points to more than 800 customers and over 10,000 robot units online by end-2022, while later 2025-2026 coverage shifts toward dual-arm deliveries and order backlog instead of refreshed enterprise- wide revenue, customer, or headcount disclosure.[CO001, CO002, CO003, CO008, CO009, CO015]

1.2 Leadership, ownership, and governance visibility

Public reporting points to a concentrated control picture rather than a fully transparent operating bench. KuCoin and 36Kr identify Chen Xi as the legal representative or chairman and place her alongside Changying Precision as a meaningful owner, while KuCoin also links her to Changying founder Chen Qixing through family ties. That combination suggests industrial parent influence, founder-linked governance, and a high degree of key-person dependence around the current transition from general industrial automation into force-controlled embodied-intelligence products. The limitation is that the public source set does not enumerate a current board roster or a broader executive bench, so investors can see the control nodes but not yet the full governance mechanics behind them.[CO012, CO016]

1.3 Funding history and valuation range

The clearest current financing fact is the May 2026 Series B/B+ raise: official and independent sources align on RMB1 billion and broadly agree on the investor set led by Hillhouse / GL Ventures and Meituan Strategic Investment, with Tencent and other investors participating. The valuation signal is less precise. EqualOcean reports an RMB8 billion post-investment value, while 36Kr, KuCoin, and Tianji's own news page frame the same event as nearly RMB10 billion. Private-market databases translate the round into roughly $1.4-$1.5 billion, but Caplight explicitly says no recent secondary-market data is available and its number is estimated from funding round information. That makes the round real, but it leaves current price discovery and exact post-money interpretation weaker than the unicorn headline.[CO004, CO005, CO006, CO007, CO010, CO011]

1.4 Milestones and remaining diligence gaps

The source set shows a coherent operating arc even though important diligence fields remain private. Official news entries trace a progression from the April 2024 launch of a force- controlled humanoid arm, to 2025 mass delivery claims for dual-arm systems, to a September 2025 strategic partnership with Zhongke Huiling, to showcase and award events in April 2026, and then to the May 2026 financing milestone. That sequence supports the thesis that Tianji has moved from legacy industrial robotics into a higher-profile embodied-intelligence position. At the same time, current public materials still do not disclose standalone 2026 revenue, gross margin, burn, runway, detailed round terms, or a fuller board roster, so later chapters should reuse the identity and milestone base here while keeping financial and governance conclusions conditional on additional diligence.[CO013, CO014, CO015, CO016, CO008, CO009]

1.5 Exhibits

2.1 Market boundary and included spend

The assigned evidence places Tianji in embodied-AI manipulation infrastructure rather than in generic AI software or a single finished-robot SKU. Current commercial scope includes industrial collaborative robots, force-controlled humanoid dual arms, joint and control modules, and related automation solutions, while newer 2026 launches start to extend the company into wheeled full-humanoid platforms and research tooling. That means the included spend today is closest to factory automation capex, humanoid OEM subsystem procurement, and research-platform equipment budgets. The sources do not support treating general-purpose AI model spend, unrelated humanoid locomotion stacks, or the entire robotics software toolchain as Tianji's current market.[CM001, CM002, CM007, CM008, CM009, CM010, CM013, CM014]

2.2 Sizing lenses are directional, not a clean TAM/SAM/SOM stack

The pool provides two useful top-down lenses and several stronger bottom-up demand proxies, but they are not directly comparable. 36Kr cites IDC for a roughly $30 billion 2026 global intelligent-robot-hardware market and cites China Development Report 2025 for a CNY400 billion China embodied-intelligence market by 2030, while multiple outlets relay Tianji-specific shipment and backlog signals that are much closer to the company's real commercial traction. Those bottom-up signals matter because they show actual buyer pull, yet they cannot by themselves define Tianji's total addressable market. For diligence, the practical read is that macro market tailwinds are large, but underwriting should rely more heavily on observable order density, customer breadth, and payload/application breadth than on any single headline TAM figure.[CM003, CM004, CM005, CM006, CM011, CM012, CM015]

2.3 Adoption drivers, buyer motion, and the main constraints

Adoption appears to be pulled by a specific bottleneck: humanoid and embodied-AI programs need reliable force-controlled manipulation hardware, not just better cognition. That creates a buyer set spanning factory automation teams, humanoid OEM platform groups, and research labs that want ready-made teleoperation and data-collection systems. At the same time, the sources repeatedly frame production consistency, yield, stability, and local support as the real gating factors for scaling from prototype programs into repeat deployments. Tianji's 2026 expansion into Gento full-humanoid products may broaden the reachable market, but it also increases execution complexity by pushing the company closer to platform and OEM exposure rather than staying purely in subsystems.[CM003, CM004, CM010, CM012, CM013, CM014, CM016]

2.4 Exhibits

3.1 Competitive boundary: supplier to humanoid OEMs, but still benchmarked against industrial robotics peers

The assigned evidence places Tianji in two overlapping competitive arenas. Independent profiles describe it as a B2B robotics company spanning industrial robots, collaborative dual-arm robots, automation solutions, and physical-AI development tooling, while official surfaces still market articulated robots, collaborative robots, controllers, software, and factory applications. That means Tianji does not compete only against other force-controlled humanoid-arm vendors; buyers can still solve similar jobs through conventional industrial automation stacks, internal integration teams, or adjacent research platforms. Caplight's closest embedding-similarity comparables are Titanium Tiger Robotics, Flexiv, and Huayan Robotics, which reinforces that outside observers still benchmark Tianji against industrial and collaborative robotics peers rather than treating it solely as a full-stack humanoid brand. At the same time, 36Kr, Embodied Global, and ManbotAI all frame Tianji's current demand around supplying humanoid OEMs and embodied-AI companies, so its real competitive field sits at the boundary between components, subsystems, and broader automation vendors.[CP001, CP002, CP003, CP004]

3.2 Capabilities favor force-control execution, but public packaging evidence is narrow

The most defensible differentiation in the retained pool is technical and operational rather than commercial. Multiple independent sources repeat Tianji's claims around force-controlled dual-arm delivery volume, while Robotopian shows a public Marvin Pro research platform with ROS 2 workflows, teleoperation, multimodal data collection, and quote-based academic bulk pricing. Those facts suggest Tianji can compete for both OEM subsystem budgets and lab/developer tooling budgets. However, public packaging evidence remains thin: the assigned sources do not expose list prices for Tianji's core OEM dual-arm supply or broader industrial arms, and they do not provide direct feature or pricing pages for Caplight's named peer set. As a result, Tianji's apparent product edge is supported better than its commercial edge, and matrix cells around pricing transparency, ecosystem lock-in, and direct like-for-like benchmarks remain materially incomplete.[CP005, CP006, CP007, CP008, CP009]

3.3 Durability depends on scale delivery and supply-chain embedment more than software lock-in

The strongest moat evidence in this pool is Tianji's combination of force-control performance claims and visible shipment/backlog scale into humanoid OEM supply chains. Independent reporting says the company delivered more than 2,000 force-controlled dual-arm units within four months in 2025 and entered 2026 with more than 10,000 units of backlog covering 45 OEM or embodied-AI customers; those are more concrete deployment markers than the prototype-stage descriptions attached to many adjacent embodied-AI entrants. But moat durability is not yet proven as software or channel lock-in. The same sources emphasize new funding for R&D, mass production, and global sales coverage, implying that execution speed and service coverage are still decisive. KuCoin's syndicated flash also situates Tianji beside other heavily financed embodied-AI names such as Unitree, Galaxy General, Looj Robotics, and Zidongbian Robotics, and EqualOcean says Tianji is incubating humanoid robot products and brands of its own. That creates a plausible future in which today's customers or adjacent OEMs become more direct competitors, while buyers still retain the fallback of conventional automation or in-house builds when Tianji's embodied-AI premium is not clearly priced or benchmarked.[CP010, CP011, CP012, CP013, CP014]

3.4 Exhibits

4.1 Revenue model and disclosed operating outputs

The retained sources describe Tianji as a B2B robotics infrastructure supplier rather than a software subscription business. Caplight classifies the company under “Hardware Sales, Services & Consulting,” while EqualOcean, Gasgoo, and 36Kr all describe a stack spanning force-controlled humanoid dual-arm systems, integrated joints, controllers, sensors, and broader automation solutions. That mix supports a revenue model centered on hardware product sales plus related systems/support work, but public disclosures stop short of showing how much revenue comes from each stream, whether services are recurring, or how revenue is recognized across equipment shipments, backlog, and any channel relationships. The traction metrics in public view are operational, not financial. Gasgoo reports more than 2,000 force-control humanoid dual arms delivered to more than 100 clients within four months of 2025, while EqualOcean and 36Kr say Q1 2026 backlog exceeded 10,000 units and covered 45 global humanoid OEMs and embodied-intelligence unicorns. 36Kr also cites older company-wide operating figures of more than 800 customers and more than 10,000 robots running online as of end-2022. Those disclosures indicate meaningful demand and installed-base breadth, but they do not reveal ASPs, acceptance timing, cancellation rates, or what portion of backlog converts into recognized revenue.[CI001, CI003, CI004, CI005, CI009, CI010, CI011, CI012]

4.2 Unit economics and cost structure

Tianji's public materials are much stronger on engineering output than on unit economics. The company and syndicated coverage consistently emphasize force-control performance, mass-production capability, automation-rate improvement, first-pass yield, and global-sales expansion, which suggests a cost structure dominated by hardware BOM, manufacturing yield, labor, and field-support costs rather than low-marginal-cost software delivery. However, none of the retained public sources disclose realized pricing, gross margin, warranty burden, channel take-rates, sales cycle length, customer-acquisition cost, payback, or working-capital intensity. That omission matters because Tianji's reported order and shipment volumes are not enough to infer revenue quality. Backlog above 10,000 units is not the same as booked revenue, and list-less hardware businesses can see wide differences between demonstration shipments, paid production deployments, pilot discounts, and full-margin recurring orders. Public evidence therefore supports only proxy analysis: Tianji appears to be scaling a capital-intensive hardware platform with meaningful production and support requirements, but the key underwriting inputs remain private.[CI002, CI004, CI005, CI009, CI010, CI011, CI012]

4.3 Capital adequacy and financing dependency

The clearest financial datapoint is the May 2026 Series B/B+ financing. EqualOcean and Gasgoo both report RMB 1 billion raised, with proceeds earmarked for R&D, large-scale mass production, and a global sales network. That use-of-funds mix implies Tianji still needs external capital to expand manufacturing capacity and overseas commercial coverage; it is not presenting itself publicly as a self-funding cash machine. The financing may be adequate for the current scale-up phase, but public evidence does not disclose cash-on-hand, monthly burn, debt, covenant load, or runway, so outside investors cannot independently test how long this round lasts. Third-party summaries also show some variance around the same financing event. InforCapital describes the May 2026 round as $138 million at a $1.4 billion valuation; EmbodiedGlobal translates it to approximately $140 million and nearly RMB 10 billion; and Caplight estimates $152.58 million total funding and a $1.5 billion valuation while saying no recent secondary-market data is available. That spread is not large enough to negate unicorn status, but it does underline that current market pricing is inferred from sparse private-market datapoints rather than transparent financial disclosure. On available evidence, the next round trigger is more likely tied to manufacturing execution and revenue conversion than to any publicly auditable profitability milestone.[CI001, CI002, CI006, CI007, CI008, CI012]

4.4 Verdict and diligence blockers

Financially, Tianji looks like a fast-scaling robotics hardware supplier with credible demand signals and a clearly funded near-term expansion plan, but the public record does not yet support a clean underwriting case on revenue quality or margin path. The positive case is that customers, shipments, and backlog all point to genuine commercial pull, and the company is using fresh capital to industrialize production and global distribution. The weak point is that every decisive financial bridge — ASP to revenue, revenue to gross profit, gross profit to burn, and burn to runway — remains undisclosed. For diligence, the highest-priority asks are product-line pricing history, realized revenue by stream, gross margin by hardware versus services, backlog aging and cancellation rates, working-capital and capex requirements, and the current cash/runway picture after the 2026 round. Until those items are produced, the best-supported conclusion is that Tianji's operational momentum is real but its valuation still outruns the public financial evidence.[CI003, CI004, CI008, CI011, CI012]

4.5 Exhibits

5.1 Portfolio breadth is broader than a single humanoid demo

The public product picture is broader than Tianji's 2026 humanoid-arm headlines. Third-party and database sources still describe a legacy industrial stack of SCARA, six-axis, and collaborative robots anchored by the self-developed Tianji Fusion control system, while Caplight classifies the business model as hardware sales plus services and consulting. That matters because it suggests the humanoid-arm line is being layered onto an existing automation supplier base rather than replacing the company's earlier industrial workflow. For product diligence, the key distinction is that Tianji now appears to serve two adjacent jobs: conventional industrial automation deployments and embodied-AI / humanoid development programs that need force control, teleoperation, and data capture.[CE001, CE002, CE009]

5.2 Externally verifiable architecture signals center on force control, joints, and developer tooling

The clearest architecture evidence comes from three different surfaces that line up with each other. Tianji's own news flow emphasizes MEMS joint torque sensors and lightweight integrated joint modules; EmbodiChain exposes a community-maintained preset with original ACD URDF naming, FK/IK tooling, and a simulation-ready configuration; and Robotopian presents Marvin Pro as a research/deployment platform with ROS 2 interfaces, multimodal data formats, and force / impedance control. Together, those sources support a product thesis built around actuation precision and deployable tooling rather than a generic humanoid concept slide. The differentiation claim is therefore strongest at the component-and-platform layer: sensors, joint modules, motion/control software, and developer interfaces that make the arm usable inside research or OEM workflows.[CE003, CE004, CE005, CE006, CE007, CE010]

5.3 Scaled shipment claims improve maturity, but manufacturing and trust controls still need direct diligence

Tianji's strongest maturity signal is that multiple sources converge on scaled force-controlled dual-arm delivery in 2025-26 rather than just prototype announcements. Official and independent coverage both point to more than 2,000 dual-arm units delivered within four months, while independent reports also describe more than 100 served customers and a Q1 2026 order book above 10,000 units across 45 global humanoid-robot OEM customers. That pushes the underwriting question from “is there a real product?” toward “how repeatable is fulfillment and support at scale?” The same independent reporting says new capital is earmarked for R&D, mass production, production-line yield, and global sales / technical support, which implies those operating capabilities are still being built out. Public product evidence in this assigned pool is also thin on named safety certifications, security attestations, uptime/status reporting, or formal quality-system disclosures for the humanoid-arm line, so trust and compliance diligence still needs direct document requests before enterprise deployment risk can be scored tightly.[CE008, CE011, CE012]

5.4 Exhibits

6.1 Installed base spans legacy factory accounts and newer humanoid buyers

The public customer picture is bifurcated. Tianji's owned web surfaces still present the company as a broad industrial robotics supplier across 3C electronics, footwear and apparel, construction machinery, and automotive manufacturing, while third-party profiles frame the business as a B2B hardware-and-services supplier rather than a pure software or research startup. Independent retellings of company data add that Tianji had more than 800 customers and over 10,000 robots online by the end of 2022, which implies the current humanoid narrative is building on a legacy industrial base rather than starting from zero.[CU001, CU002, CU003, CU004, CU011]

6.2 Humanoid adoption claims are stronger on volume and cohort reach than on named accounts

Tianji has more concrete proof of adoption than many private robotics companies, but the proof is uneven. The strongest public datapoint is a company-originated claim, echoed by 36Kr, that Tianji delivered more than 2,000 force-controlled humanoid dual-arm units within four months during 2025. Third-party coverage also repeats that first-quarter 2026 on-hand orders exceeded 10,000 units and that the customer set covered 45 global humanoid robot OEMs and embodied-intelligence unicorn enterprises. Those figures point to real pipeline depth, yet outside one named Zhongke Huiling partnership and order, the public record still exposes far more cohort-level proof than individually referenceable production deployments.[CU005, CU006, CU007, CU010, CU014]

6.3 Expansion appears to rely on customer enablement as much as on hardware specs

The adoption loop Tianji is describing is not just sell-a-robot and leave. Official support pages advertise phone, remote, and on-site fault support, spare-parts planning, software guidance, and training delivered by robotics engineers, while a third-party Marvin Pro product listing adds product-level warranty terms and academic discounts for universities and research institutions. Separately, 36Kr and EqualOcean both say Tianji plans to expand localized sales and technical-support teams overseas, especially in North America, on top of its existing OEM customer coverage. That combination suggests land-and-expand ambitions, but the public sources do not quantify attach rates, renewal behavior, or service revenue.[CU008, CU009]

6.4 Durability and concentration are still the weakest parts of the customer case

The biggest diligence gap is not whether Tianji has customer logos or even whether it ships product; it is whether the installed base renews, expands, and diversifies economically. The assigned pool does not disclose NRR, GRR, churn, contract duration, backlog conversion, top-customer exposure, channel dependence, or backlog aging by product line. Even the lone explicitly risk-framed third-party artifact retained for this chapter only surfaces as an aVenture title rather than as usable underlying analysis. That leaves the chapter able to say Tianji is commercially active, but not able to underwrite customer durability with the same confidence as adoption momentum.[CU012, CU013]

6.5 Exhibits

7.1 Ranked risk picture

The highest-confidence public risks are execution and pricing risks rather than proof of an outright product or financing fabrication problem. Multiple sources align that Tianji raised roughly RMB1 billion in May 2026 and then added a September 2026 B++ round, while claiming more than 2,000 force-controlled dual-arm deliveries in four months and more than 10,000 units on order by Q1 2026. Those same sources also show that the new capital is still being aimed at manufacturing automation, yield improvement, and localized overseas support, which means the operating model is still being hardened while investors are already being asked to price the business at unicorn-scale marks. Public evidence therefore supports a risk ranking led by valuation/disclosure opacity, manufacturing scale-up, customer/OEM concentration, and overseas support/compliance execution.[CR001, CR002, CR003, CR004, CR005, CR007, CR015]

7.2 Governance, regulatory, and financial disclosure risk

The sharpest financial underwriting risk is not whether Tianji raised money, but whether an investor can underwrite the price and control surface with precision. Public sources point to a May 2026 post-money valuation anywhere from RMB8 billion to nearly RMB10 billion, while market-data pages translate that into roughly $1.4 billion to $1.5 billion. The same public record is also inconsistent on ownership language: Robot Belt and 36Kr describe Changying or Everwin retaining about a 27 percent stake, whereas EqualOcean still describes Tianji as a wholly owned subsidiary. That does not prove misconduct, but it does mean minority investors should treat cap-table precision, board rights, preference stack, and any secondary component as unresolved diligence items rather than assuming the public story is already clean. Public materials in the assigned pool also do not surface named export-control, certification, product-safety, or litigation disclosures commensurate with Tianji's stated North America expansion ambitions, so regulatory readiness should be treated as unverified rather than presumed.[CR002, CR003, CR007, CR009, CR010, CR011]

7.3 Operational, partner, and customer execution risk

Operationally, Tianji has meaningful momentum but also a narrow path to proving that momentum is durable. Independent and company-linked sources repeatedly cite more than 2,000 dual-arm units delivered in four months and more than 10,000 units on order by Q1 2026, but the same evidence makes clear that management is still investing new capital to raise automation, improve first-pass yield, and stand up localized support teams abroad. That is consistent with a company that may have found product-market pull but has not yet publicly demonstrated mature service, reliability, or backlog-conversion metrics. Customer and partner concentration add to that risk: the disclosed backlog is tied to 45 global humanoid OEMs and embodied-AI companies, the product listing reviewed here shows a limited hardware warranty footprint, and public evidence still does not disclose cancellation rates, top-customer exposure, renewal behavior, or warranty-claim performance. The result is high residual execution risk even if product demand is real.[CR004, CR005, CR006, CR007, CR008, CR012, CR013, CR014, CR015]

7.4 Mitigations, monitoring indicators, and kill criteria

The current mitigants are tangible but incomplete: Tianji has raised additional capital, continues to invest in core torque-sensor and control IP, and explicitly plans localized sales and technical-support teams to reduce overseas execution risk. Those steps matter, but they are still plans or early build-out indicators rather than disclosed proof that quality, field support, and customer economics are already de-risked. The practical monitoring set for the next diligence cycle is therefore narrow and measurable: whether the company can reconcile the May 2026 valuation range and ownership disclosures; whether backlog converts into shipped, accepted, and repeat orders; whether overseas support entities, certifications, and channel partners become visible; and whether field-reliability metrics or large-customer references become public. A thesis-break would be any credible evidence of a down-round versus the May mark, major backlog slippage, material safety or compliance issues, or customer concentration meaningfully worse than the current 45-OEM framing implies.[CR002, CR005, CR007, CR011, CR012, CR015]

7.5 Exhibits

8.1 Underwriting call

Tianji's May 2026 round appears real and strategically important: the company and independent coverage align on RMB 1 billion of new capital, while third-party market-data vendors cluster the financing around roughly $1.4-1.5 billion post-money. That is enough to treat the latest round as a credible reference point, but not enough to treat it as a durable bargain. The same public package that supports the mark also leaves the crucial bridge from backlog and shipment claims to revenue, gross margin, cash burn, and repeat purchasing completely undisclosed. With real traction signals but an incomplete operating-value bridge, the price-sensitive call is track rather than buy, and entry discipline should stay at or below the last round until management opens the financial and term data room. [CV001, CV003, CV005, CV007, CV009, CV012, CV013]

8.2 Price support and public mark quality

Public valuation support is directionally consistent but not clean. Multiples, Caplight, and InforCapital all anchor Tianji near $1.4-1.5 billion, while the company-led English page and 36Kr frame the round at nearly RMB 10 billion post-money. EqualOcean, however, prints the same RMB 1 billion financing at RMB 8 billion post-investment, which is a meaningful gap rather than a rounding error. Caplight also reports no recent secondary market data and no public IPO plans, limiting any external read-through on present liquidity or price discovery. The result is a usable last-round mark but not a well-corroborated premium valuation case, especially because the only public term detail in the assigned pool is Multiples' note that the round used convertible preferred shares. [CV001, CV002, CV003, CV004, CV005, CV006, CV010, CV011, CV012]

8.3 Scenario framing

The upside case rests on commercialization conversion, not on storytelling about embodied-intelligence demand. Tianji has enough product and delivery evidence to justify continued diligence: public sources describe more than 2,000 dual-arm units delivered in 2025, more than 10,000 units of Q1 2026 orders, broad OEM coverage, and a legacy installed base across industrial customers. But those claims remain a demand proxy until they are translated into disclosed revenue, gross margin, backlog quality, and customer concentration data. Accordingly, the base case is a hold-the-mark scenario around the latest financing band, the bull case requires proof that backlog converts into quality revenue, and the bear case is driven by opaque terms or weaker-than-advertised conversion. [CV007, CV008, CV009, CV012, CV013, CV014]

8.4 Exit readiness and final diligence asks

Public evidence does not yet support an exit-underwriting narrative beyond another private financing or a strategic outcome tied to Tianji's position in the humanoid-robot supply chain. Caplight's page explicitly says no public IPO plans have been announced and that no recent secondary market data is available, which makes the public market exit window speculative rather than evidenced. The minimum diligence package before paying into or above the latest round is straightforward: verify revenue and gross margin conversion from the claimed order book, inspect liquidation preference and anti-dilution terms, test repeat buying and concentration across the stated customer base, and confirm whether secondary liquidity exists at or above the headline mark. Absent those proofs, the thesis stays interesting but not yet underwritable at a premium. [CV003, CV004, CV009, CV010, CV014, CV015]

8.5 Exhibits