2026 has been repeatedly defined as the "Year One of Mass Production" for humanoid robots. While predictions for total unit shipments remain inconsistent, capacity planning in the joint module sector has already surged ahead. Placing these figures side by side reveals a tension: supply-side confidence is high, but does the demand-side ledger have room for every plan?
01、Three Magnitudes of Capacity Figures
According to recent reports, capacity deployment in the joint sector has unfolded across three magnitudes:
Order Level. Zhuoyu Technology secured bulk procurement orders from national-level customers for 100,000 planetary joint modules, with total order value exceeding CNY 100 million. This is one of the largest single disclosed orders in the joint field to date; the customer is described as a "national-level client," and the nature of the order is bulk procurement rather than a framework intention.
Production Line Level. Fulin Precision—a listed company primarily engaged in automotive engine components—announced an investment of CNY 500 million to build a robotic intelligent electric joint project. Phase I fixed asset investment is approximately CNY 200 million, with a planned annual output of 2 million sets of intelligent electric joints, expected to commence production within nine months. Meanwhile, Ant Factory (Mayi Gongchang), in partnership with Quanzhibo and LinkerBot, established a joint venture named Beijing Zhiling Factory, planning an annual output of 1 million joint modules.
Monthly Level. Quanzhibo's monthly production capacity surpassed 60,000 units in the second half of 2026, annualizing to approximately 720,000 units; LinkerBot shipped over 4,000 units per month, holding more than 80% of the global market share for high-degree-of-freedom dexterous hands.
Zhuoyu's 100,000 units represent signed delivery commitments, while Fulin's2 million sets and Zhiling Factory's 1 million are production line plans. Quanzhibo's 60,000 units/month reflects current actual output capability. The substance of these three categories differs significantly, yet they are often conflated in narratives about "joint capacity."
02、Calculating the Demand Side
Vice Minister Ke Jixin of the Ministry of Industry and Information Technology (MIIT) disclosed at a State Council Information Office press conference that China's humanoid robot production reached approximately 20,000 units in 2025, exceeded 40,000 units in the first half of 2026, and is projected to surpass 100,000 units for the full year. Assuming each humanoid robot requires an average of 20–30 joint modules, the corresponding demand for joint modules is approximately 2–3 million units.
Using the most optimistic estimate of 100,000 humanoid robot units produced, supply-side plans have already touched the demand ceiling: Fulin Precision's Phase I capacity of 2 million sets, combined with the 1 million units planned by the Ant Factory-affiliated joint venture, totals exactly 3 million units, reaching the upper limit of the annual demand forecast. If we further include Quanzhibo's annualized output capability of approximately 720,000 units, Zhuoyu's signed orders of 100,000 units, and joint demands from dexterous hand manufacturers like LinkerBot, the total planned supply clearly exceeds the demand cap.
Viewed from another angle, IDC data indicates that global humanoid robot shipments in 2026 are expected to exceed only 50,000 units. The gap between production volume and shipment volume points to a critical fact: production lines are preparing for 100,000-unit scale output, but the global market may actually absorb only half of that figure.
The conclusion is: there is no long-term window of "undersupply" in terms of total volume for the joint sector. The logic behind capacity planning is not "produce what the market needs," but rather "secure the production lines first, betting that competitors cannot produce or cannot produce cheaper." This is a competition using capital to secure positions, not a sharing of incremental growth.
03、The Substance Behind Capacity Figures
Even among "million-level capacities," the substance can vary across three tiers:
Tier 1: Capacity Anchored by Customers. Zhuoyu Technology's 100,000-unit order comes from national-level bulk procurement, totaling over CNY100 million. The order itself anchors the delivery schedule and acceptance standards. This represents a real mapping of demand, making the expansion logic valid.
Tier 2: Capacity Verified by Production Lines. Quanzhibo's breakthrough of 60,000 units per month is an operational output capability currently in effect, not just a planned figure. Zhuoyu Technology disclosed a "product delivery yield rate exceeding 95%," and has obtained IATF 16949 automotive quality management system certification and ASIL-B functional safety certification. Only when there is yield data, certification, and actual shipments does it constitute verifiable capacity.
Tier 3: Capacity That Is Merely Planned. Fulin Precision's Phase I capacity of 2 million sets explicitly states in the announcement that "interior decoration will begin within three months after factory handover, with production expected to start within nine months," and that investment will be dynamically adjusted in two phases. This means this capacity will contribute to revenue only after production begins, and Phase II scale will be "dynamically adjusted based on market demand"—the plan itself leaves an exit route. Whether the capacity can be absorbed by market demand remains uncertain.
Separating these three tiers reduces the perceived intensity of the "arms race." The number of players who have truly entered Tier 2 or above is far fewer than the impression given by the capacity figures.
04、The True Source of the Bubble
More alarming for joint suppliers than "total oversupply" are two structural cracks on the demand side.
First, leading OEMs are shifting joints from "externally sourced components" to "in-house developed" or "capital-bound" components.
In Eclobot's financing round exceeding CNY 100 million, Ecovacs (existing shareholders doubling down), X Square Robot (OEM), and Dechak Electric Motor (automotive motor supplier) appeared simultaneously in the investor list. Industrial capital is no longer satisfied with mere supply relationships but is locking joint suppliers into their own supply chain maps through equity stakes. For invested companies, this is an endorsement; but for unbound third-party joint companies, it means: the addressable external sourcing market share is being carved up in advance by equity relationships.
Second, cross-industry entrants are blurring the boundary between "supplier" and "competitor."
Fulin Precision partnered with AgiBot in late 2024 to enter the joint component supply chain, and in March 2026 established a subsidiary specifically dedicated to R&D and manufacturing of electric joints, covering planetary, harmonic, and cycloidal routes. An auto parts enterprise entering the sector with CNY 500 million and a 2-million-set capacity plan derives its advantage not from accumulated joint technology, but from the transfer of precision manufacturing capabilities and cost control systems. When such players bring the production discipline of the automotive industry into the joint sector, pure robotics startups face rapidly increasing pressure on costs.
05、The Real Watershed Is Not Capacity Size
If capacity is not a moat, what is?
Based on clues from recent reports, the watershed lies in three areas:
1. Verifiability of Delivery. The reference value of Zhuoyu Technology lies not in the number "100,000" itself, but in its simultaneous disclosure of three verifiable metrics: yield rate exceeding 95%, ASIL-B functional safety certification, and IATF 16949 automotive-grade system. Once a joint's high output torque fails, it becomes a personal safety issue. Automotive-grade certifications are transitioning from bonus points to entry barriers. The time barrier constituted by certifications and yield rates is harder to catch up with in the short term than capacity figures.
2. System Matching Capability Rather Than Single-Point Parameters. Eclobot's strategy extends from joints to underlying layers, having formed a technical system covering seven core components: reducers, motors, bearings, encoders, drivers, force sensors, and brakes. Fulin Precision is simultaneously deploying planetary, harmonic, and cycloidal routes. Excelling at a single joint is insufficient; cross-route, cross-component system integration capability is becoming a core consideration for OEMs when selecting partners.
3. Thermal Management—The Invisible Veto Item Determining "Operational Availability." Coolink Robotics raised tens of millions in angel funding with thermal management of joints as its differentiation direction. Founder Yuan Yanjie judged that while the early industry focused on torque, volume, and response speed, once robots enter real-world scenarios, "continuous working capability cannot be ignored." Zhuoyu Technology treats full-link temperature control as the core support for mass production delivery—motor efficiency at 92%, GaN driver efficiency exceeding 95%, and single-stage planetary reducer efficiency at 97%—maintaining thermal balance by reducing heat sources rather than post-hoc dissipation. In the eyes of factory clients, a robot that frequently stops due to overheating is non-compliant equipment. Thermal management capability directly links to MTBF (Mean Time Between Failures) and O&M costs, the latter being a core variable in bulk procurement decisions.
Conclusion
The capacity race in the joint sector is essentially a positioning battle exchanging capital for time. Fulin Precision's 2-million-set plan, the Ant Factory-affiliated1-million-unit joint venture capacity, Quanzhibo's actual output of 60,000 units/month, and Zhuoyu's signed orders of 100,000 units—piecing these numbers together depicts not a market of "undersupply," but a sector where supply-side planning has hit the ceiling, actual global shipments may be only half of the planned production, and structural divergence is accelerating.
What will truly be remembered in 2026 will not be which company announced the largest capacity figure, but which company turned capacity into good products delivered on time, and whose customers are still willing to place orders the following year.
This article is compiled based on publicly available online information, for reference only, and does not constitute investment advice.