Why have deployment expectations been pushed back despite continued strong interest in adoption?

  • Of the 94 Chinese companies surveyed by Morgan Stanley, 64% plan to conduct pilots or major projects involving wheeled or legged humanoid robots over the next three years, slightly above 62% in 2025. But the shares expecting to deploy wheeled and legged robots by the end of 2027 fell from last year’s 58% and 41% to 36% and 22%, respectively. The bank estimates that the adoption curve has been pushed back by roughly a year.
  • The slight increase in adoption interest also reflects a change in the sample: manufacturers’ share of respondents rose from 51% to 65% this year, and manufacturers are more likely to adopt robots early. All respondents came from companies with annual revenue of at least 1 billion yuan, so the survey reflects the purchasing plans of such companies.
  • Actual use remains at an early stage: 10% of companies have launched pilots or major projects for wheeled robots and 6% for legged robots, compared with 6% and 5%, respectively, last year. Even by the end of 2028, companies expecting to deploy 1–10 units remain the largest group. The industry is closer to broadening the coverage of trials than rapidly expanding individual fleets.

What economic hurdles must be cleared to move from pilots to bulk purchases?

  • Among 60 potential adopting companies, 72% believe a unit price below 100 thousand yuan is needed for widespread use, up from 53% last year; 82% require payback within three years, and 65% require payback at least as fast as for other automation investments. Purchasing decisions are shifting from demonstrations of functionality to whether robots can generate sufficient returns.
  • Budget attitudes are also more cautious: the share expecting a significant increase in robot investment over the next three years fell from 30% to 22%. High upfront costs became the leading barrier to adoption, cited by 62%; satisfaction with existing products fell from 23% to 18%, while dissatisfaction rose from 25% to 28%. Cost pressures and inadequate performance are both slowing adoption.
  • A low selling price is not the only solution. Morgan Stanley believes high-value production tasks can support higher prices, but robots must deliver measurable productivity gains, maintain high utilization and keep system integration and maintenance costs under control.

Which use cases are most likely to generate sustainable demand first?

  • With industries weighted equally, potential adopters’ intentions to deploy robots in manufacturing and logistics and warehousing were 90% and 85%, respectively. Output, safety and labor savings are easier to measure in repetitive processes such as material handling, loading and unloading, packaging and palletizing, making these better suited to establishing commercial value first.
  • Morgan Stanley expects mobile manipulators and wheeled robots to enter industrial use before bipedal models: structured factories and warehouses generally do not need complex walking mechanisms. Legged robots are better suited to unstructured environments and high-value tasks, but face more demanding technical and reliability constraints.
  • The bank forecasts China’s humanoid robot shipments at 50 thousand units in 2026 and 100 thousand units in 2027, and expects pilots for certain specific tasks to expand deployment starting in 2027. More meaningful indicators of progress are repeat orders, deployment across factories and fleets of more than 10 units.

With robot shipments scaling slowly, why might components benefit earlier?

  • Morgan Stanley’s reasoning is that robot makers have increased pilots to gather data from real-world settings, already generating component purchases, so suppliers’ revenue opportunities may arise before large-scale deployment by end users. The bank is positive on the long-term robotics trend and favors Hengli Hydraulics, Leaderdrive and Shuanghuan Driveline.
  • Leaderdrive offers concrete order evidence: orders for humanoid robot-related reducers were about 400 thousand units in the first nine months of 2026, versus about 100 thousand units for all of 2025. The company aims to raise monthly capacity from about 90 thousand units currently to 120 thousand units by year-end. But rapid capacity expansion and price competition could squeeze margins, and costs remain a constraint on translating order growth into improved profitability.