China's Better-Than-Expected Mass Production Drives Humanoid Robot Forecast Upgrade, with the Ecosystem Flywheel Entering an Acceleration Phase
AI summary card
China's Better-Than-Expected Mass Production Drives Humanoid Robot Forecast Upgrade, with the Ecosystem Flywheel Entering an Acceleration Phase
Deutsche Bank raises its 2026 global humanoid robot shipment forecast from approximately 50,000 units to approximately 80,000 units, and believes data expansion, model iteration, and cost declines could drive an industry breakthrough before 2030.
- China's humanoid robot production in the first half of 2026 has exceeded 40,000 units, significantly faster than previously expected.
- The 2026 global shipment forecast is raised to approximately 80,000 units, with China expected to contribute approximately 70,000 units.
- The 2030 global market forecast is raised to 1 million units and US$25bn, and the 2050 forecast is raised to 100 million units and US$1.5tn.
- Task success rates in industrial scenarios range from below 50% to over 99%, while typical payback periods still exceed two years.
- The U.S. restricts advanced robotic equipment produced abroad, benefiting U.S. manufacturers, having a neutral impact on Japan, and creating divergent impacts for Chinese manufacturers.
Report interpretation
Overview
The report reviews global humanoid robot industry progress in the first half of 2026. Chinese manufacturers rapidly expanded production, Tesla and Figure in the U.S. began increasing manufacturing capabilities, and Europe, Japan, and South Korea were also driven by localization demand. Based on China's better-than-expected production progress, the report broadly raises near- and long-term market forecasts, while emphasizing that industrial applications have not yet formed a stable, widely replicable business model.
Core views
Industry growth is shifting from single-point technological breakthroughs to ecosystem expansion driven by mutual reinforcement among hardware, models, data, and applications. China will remain the largest market in the short term, contributing approximately 70,000 of the approximately 80,000 global shipments expected in 2026; over the long term, U.S. shipment scale is expected to gradually approach China's. Entertainment performances and academic research already have viable commercial cases, while industrial scenarios remain constrained by success rates, costs, and payback periods. However, continued data accumulation, model evolution, and cost reduction could enable the industry to cross the commercialization inflection point before 2030.
Analysis framework
The report combines corporate production and capacity targets, government department forecasts, channel checks, manufacturer production ramp-up information, and real-world task performance to build a bottom-up regional and company-level breakdown of global shipments. It also uses the four elements of hardware, models, data, and applications to explain the positive feedback mechanism of the industry ecosystem; equity mapping focuses on component value per unit, customer expansion, vertical integration capabilities, and commercialization timelines.
Methodology notes
Hardware, Models, Data, Applications
Better hardware expands application boundaries, applications generate more real-world data, data drives model iteration, and improved model capabilities in turn enhance task success rates and hardware demand, thereby forming a self-reinforcing cycle.
Corporate Capacity and Regional Demand Breakdown
The global shipment forecast is aggregated based on actual production by Chinese manufacturers, production ramp-up by U.S. manufacturers, companies' long-term capacity plans, and localization demand across regions.
Dual Validation of Technical Reliability and Economics
Application readiness for scaled deployment is assessed through real-world industrial task success rates, equipment costs, and investment payback periods, rather than evaluating technology maturity solely based on demonstrations.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Hengli Hydraulic(601100.SS)A beneficiary in humanoid robot components such as screws and motors, rated BUY with a target price of RMB124.
- Strengths
- Value per unit is increasing, with coverage of body screws, body motors, and hand motors, while customer expansion continues; the construction machinery business provides cyclical support.
- Weaknesses
- Robot component revenue in 2026 remains relatively small, and growth depends on customers' smooth production expansion.
- Comparison
- Compared with single-component suppliers, it has broader product coverage and diversified opportunities in construction machinery, bionic limbs, aerospace, and machine tools.
- Risks
- Customer mass-production delays, component validation falling short of expectations, and industry price declines could suppress revenue realization.
- Shuanghuan Driveline(002472.SZ)A potential beneficiary in reducers for high-load rotary actuators, rated BUY with a target price of RMB47.
- Strengths
- Has conducted robot R&D cooperation with Tesla for more than three years and has growth drivers in new businesses such as traditional robots, industrial equipment, and intelligent actuators.
- Weaknesses
- The core automotive gearbox business faces pressure in 2026, and robot-related orders still need to be converted.
- Comparison
- Combines a foundation in automotive transmission manufacturing, accumulated robot reducer R&D capabilities, and overseas expansion potential.
- Risks
- Changes in Tesla's technology roadmap, R&D results not converting into orders, and declines in the core business could all affect performance.
- Harmonic Drive Systems(6324.T)A core supplier of reducers and actuators for humanoid robots outside China, rated BUY with a target price of ¥9,900.
- Strengths
- After expanding from reducers to complete actuators, the potential average selling price is approximately three times that of standalone reducers; aerospace and semiconductor businesses also maintain relatively fast growth.
- Weaknesses
- Volume ramp-up of robot actuators remains constrained by end-market mass-production timing.
- Comparison
- It has strong reducer technology and customer positioning in supply chains outside China.
- Risks
- Delayed end demand, intensifying competition, and customer in-house development could reduce market share and profitability.
- Tesla(TSLA.US)Optimus vertical integration platform, rated BUY with a target price of US$420.
- Strengths
- It has hardware, manufacturing, computing power, model training, and internal factory application scenarios, and can continuously accumulate data through internal deployments; hand hardware bottlenecks are expected to ease.
- Weaknesses
- The early manufacturing ramp is expected to be slow, and general task capabilities and large-scale mass production have not yet been fully validated.
- Comparison
- Compared with most startups, Tesla has a more complete manufacturing system, computing resources, and internal application closed loop.
- Risks
- Optimus mass-production progress, technical reliability, or cost improvements may fall short of expectations; external sales are expected to begin only in late 2027 or early 2028.
- Mobileye(MBLY.US)Entering the warehousing and commercial humanoid robot market through Mentee Robotics, rated BUY with a target price of US$13.
- Strengths
- Aims to provide relatively low-cost products, with a bill of materials cost of approximately US$50k at low production volumes and a path to reduce this by 50% after scale increases; it has already conducted proof-of-concept projects with multiple customers.
- Weaknesses
- The business remains in the early validation stage, and reaching a scale of several thousand units may not occur until 2028 to 2029.
- Comparison
- The market may not fully reflect the value of the Mentee business, with upside potential over the next 9 to 12 months.
- Risks
- Proof-of-concept projects may fail to convert into commercial orders, cost reduction progress may fall short of expectations, and there are M&A integration risks.
Key data
- 2026 Global Shipment ForecastApproximately 80,000 unitsPrevious forecast was approximately 50,000 units; the chart corresponds to a forecast value of 79,000 units.
- 2026 China Shipment ForecastApproximately 70,000 unitsChina is expected to continue holding the largest global market share.
- China's Production in the First Half of 2026Over 40,000 unitsActual production progress significantly exceeded the report's previous expectations.
- Unitree Cumulative ProductionApproximately 11,000 units as of May 2026Only bipedal humanoid robots are counted.
- AGIBOT Cumulative ProductionApproximately 15,000 units as of June 2026Includes bipedal and wheeled embodied AI robots.
- 2030 Global Market Forecast1 million units, US$25bnThe previous shipment forecast was 700,000 units.
- 2050 Global Market Forecast100 million units, US$1.5tnThe previous shipment forecast was 70 million units.
- Task Success Rates in Industrial ScenariosBelow 50% to over 99%There are significant differences across tasks and deployment conditions.
- Typical Payback Period for Industrial ApplicationsOver two yearsMost current industrial deployments have not yet reached sufficient commercial viability.
- Figure Production SpeedIncreased from 1 unit per day to 1 unit per hourProduction in April 2026 was approximately 150 units.
Impact & implications
The forecast upgrade reinforces the medium- to long-term demand logic for robot body, screw, motor, reducer, and actuator suppliers, and enhances the strategic value of companies with full-stack capabilities or advantages in core components. Near-term investment opportunities are more likely to concentrate in companies with existing customer validation, rising value per unit, and clearer capacity delivery, rather than those relying solely on long-term themes. U.S. policy could drive local supply chain and local manufacturing investment, while increasing uncertainty for Chinese manufacturers entering the U.S. market.
Risks
- Task success rates in industrial scenarios vary greatly, and current robots still cannot reliably perform a broad range of general-purpose tasks.
- Typical investment payback periods exceed two years; if cost declines or utilization improvements are slower than expected, the commercialization inflection point may be delayed.
- Manufacturers such as Tesla are in the slow early phase of the manufacturing S-curve, and public capacity targets are not equivalent to actual output.
- Long-term market forecasts depend heavily on multiple assumptions, including data scale, model capabilities, cost curves, and application demand.
- U.S. restrictions on advanced robotic equipment produced abroad are broad in scope and could curb imports of new products from Chinese manufacturers and their expansion into the U.S. market.
- Component suppliers face risks from changes in customer technology roadmaps, in-house substitution, price competition, and order concentration.
What to watch
- Whether China's full-year 2026 production can approach MIIT's forecast of 100,000 units.
- Whether Unitree and AGIBOT can respectively achieve annual targets of 10,000 to 20,000 units and 20,000 to 40,000 units.
- Whether Tesla Optimus can reach channel expectations of 1,000 units per week in the third quarter of 2026 and over 2,000 units per week by year-end.
- Whether Tesla expands internal deployment in the fourth quarter of 2026 as planned and launches external sales in late 2027 to early 2028.
- Figure's progress in converting a production speed of 1 unit per hour into stable batch deliveries.
- Whether industrial task success rates, unit costs, and payback periods can continue to improve.
- Implementation details of U.S. advanced robot import restrictions and their actual impact on Chinese manufacturers such as Unitree and Geekplus.
- Robot orders and value-per-unit realization for Hengli Hydraulic, Shuanghuan Driveline, and Harmonic Drive Systems.
- Customer proof-of-concept and commercial order progress for Mobileye's Mentee over the next 9 to 12 months.