Humanoid robots face not only a technology problem, but also a “PR and social license” problem

Overview

This report updates Morgan Stanley’s latest views on the global humanoid robotics ecosystem. Its core judgment is that the humanoid robotics industry is moving from product demonstrations toward real deployment in structured scenarios, particularly in manufacturing, logistics, inspection, public services, and certain service-industry applications. At the same time, the report cautions that the industry’s main bottlenecks are not limited to robot hardware, AI models, or manufacturing costs; they also include social license, labor relations, policy restrictions, commercial ROI, and verifiable operating data.

Core views

The report believes the long-term opportunity for humanoid robotics remains enormous, but short-term commercialization is still at an early stage. If the industry narrative overemphasizes “replacing workers with robots,” it may increase resistance from unions, employees, and regulators. By contrast, positioning robots as tools that augment human work, alleviate labor shortages, and handle dangerous or repetitive tasks may make deployment approval easier. From a technology perspective, robot AI architectures are still evolving rapidly. Ultimately, the advantage may come not from a single model, but from real-world deployment data, rapid learning loops, and deep integration of hardware and AI software. In China, policy, state-owned enterprises, and local governments are promoting validation in high-value scenarios, prompting Morgan Stanley to raise its 2026 forecast for Chinese humanoid robot shipments.

Analysis framework

The report combines value-chain tracking, policy analysis, company-event reviews, financing and patent data, equity-basket performance, and long-term TAM estimates. Its focus is not on providing financial forecasts for a single company, but on determining which factors will influence adoption speed, the order in which supply-chain participants benefit, and the validation framework used by investors as humanoid robots move from proof of concept toward commercial deployment.

Methodology notes

• Industry adoption frameworkTechnology diffusion and commercial ROI validation

  Adoption barriers from demonstration to deployment

  The report believes that actual operating hours, intervention rates, maintenance costs, repeat-purchase rates, customer payback periods, and real-world output provide stronger evidence of commercialization progress than demonstration videos, theoretical capacity, and non-binding orders.

• Policy and geopolitical riskGUARD Act and import restriction scenario

  Robot trade and security review

  The proposed US GUARD Act could require national security agencies to review humanoid and quadruped robots produced in China and other countries. If the review is not passed or is not completed within one year, the products could be added to the FCC Covered List and restricted from entering the US market.

• Value-chain mappingHumanoid 100 and the Chinese humanoid robotics value chain

  Tracking global and Chinese supply-chain beneficiaries

  The report tracks the performance of system manufacturers, robot brains, actuators, batteries, bearings, sensors, LiDAR, reducers, ball screws, and integrators through the Humanoid 100 and Chinese humanoid robotics value-chain lists.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Tesla Optimus

  Representative of robot-system manufacturing and long-term capacity building

  Strengths

  According to the report, Tesla is converting its Fremont production line for Optimus, supporting a narrative centered on manufacturing engineering and scalability.

  Weaknesses

  The report cautions that long-term designed capacity does not equal near-term shipments, yield, or real demand.

  Comparison

  Compared with XPeng and Boston Dynamics, Tesla places greater emphasis on vertical integration and large-scale manufacturing potential.

  Risks

  Actual mass-production timing, customer ROI, operating reliability, and commercial implementation remain to be validated.

• XPeng IRON

  Representative of a Chinese automaker entering humanoid robot manufacturing

  Strengths

  According to the report, XPeng is targeting monthly IRON capacity of more than 1,000 units within the year and plans to deploy the robots first in its own stores.

  Weaknesses

  Current information is closer to a capacity target than confirmed sales or deliveries.

  Comparison

  Like Tesla, it has an automotive manufacturing background, but in the near term it is more focused on validation through Chinese scenarios and its own channels.

  Risks

  International launches, commercial orders, application efficiency, and cost recovery remain uncertain.

• Boston Dynamics / Atlas

  Representative of advanced robotics technology and labor-relations issues

  Strengths

  Boston Dynamics has deep robotics R&D capabilities and is planning a new manufacturing and AI center.

  Weaknesses

  Hyundai’s union has proposed restrictions on Atlas entering production lines, showing that social license may constrain deployment.

  Comparison

  Compared with emerging robot-system companies, Boston Dynamics has stronger technological experience, but its commercialization pace and labor coordination are subject to greater scrutiny.

  Risks

  Union negotiations, employee retraining, deployment agreements, and safety certification could affect the pace of implementation.

• Hesai HSAI.US

  Potential beneficiary in the humanoid robot perception and LiDAR supply chain

  Strengths

  The report added Hesai to Humanoid 100, highlighting its global LiDAR manufacturing capabilities, ADAS customer pipeline, and opportunities related to humanoid robot actuator modules.

  Weaknesses

  Humanoid robot-related revenue remains at an early stage, and its contribution needs to be validated over time.

  Comparison

  Compared with pure-play robot-system companies, Hesai operates in upstream perception and components and is exposed to demand from multiple robotics and autonomous-driving categories.

  Risks

  Customer ramp-up timing, pricing pressure, gross-margin realization, and policy restrictions could affect earnings.

• Unitree

  Representative of a low-cost Chinese robot platform and research ecosystem

  Strengths

  The report states that Chinese-manufactured robots, particularly Unitree products, are widely used by US research laboratories because of their low cost and high availability.

  Weaknesses

  If the US restricts imports of Chinese robots or component cooperation, overseas research and commercial expansion could be impeded.

  Comparison

  Compared with US and European robot-system manufacturers, Unitree stands out in cost and supply availability.

  Risks

  The US GUARD Act, cybersecurity reviews, import restrictions, and geopolitical risks.

Key data

• 2026 forecast for Chinese humanoid robot shipments50,000 unitsMorgan Stanley raised its forecast from 28,000 units to 50,000 units.
• 2030 forecast for Chinese humanoid robot shipments446,000 unitsThe report believes integrators will use deployment scale to create data feedback loops.
• Chinese policy targetDeploy 10,000 humanoid robots by the end of 2026 and establish more than 100 high-value applicationsThe target covers manufacturing, logistics, healthcare, and disaster response scenarios, but policy targets do not necessarily equate to sustainable commercial demand.
• Global adoption forecastApproximately 28.1 million units in 2036, approximately 138.5 million units in 2040, approximately 430 million units in 2044, and approximately 1 billion units in 2050These are estimates of the long-term adoption path presented in the report.
• Humanoid 100 performanceUp 44% since its establishment on 2025-02-06On an equal-weighted basis, it outperformed the S&P 500, MSCI Europe, and MSCI China, but lagged MSCI Korea and MSCI Taiwan.
• July MTD performance of the Chinese humanoid robotics value chainDown 11%On an equal-weighted basis, it underperformed MSCI China’s 7% decline, reflecting weakness in the AI sector and subdued sentiment while awaiting mass-production updates from leading integrators.
• New addition to Humanoid 100Hesai HSAI.USHesai was added to Humanoid 100 due to its LiDAR strengths, ADAS customer pipeline, and opportunities related to actuator modules for humanoid robots.
• Removal from Humanoid 100ORCLThe report stated that it was removed because of insufficient relevance to the relative investment story.

Impact & implications

The investment implication is that humanoid robots remain a compelling long-term theme, but in the short term the focus needs to shift from conceptual excitement to verifiable commercial metrics. Supply-chain companies may benefit from capacity building, prototype iteration, and policy support, but sustained re-rating will require evidence of customer repeat purchases, stable operations, declining costs, and expansion into application scenarios. US protectionism could protect domestic robot manufacturers, but it could also increase R&D costs because US research laboratories extensively use low-cost, readily available Chinese robots. The Chinese market is being strongly supported by policy and state-owned enterprise applications, but the sustainability of its economic demand still needs to be validated.

Risks

• Labor-relations and public-narrative risk: If robots are viewed as directly replacing workers, deployment may face backlash from unions, employees, and regulators.
• Policy and geopolitical risk: If the US GUARD Act is enacted, it could restrict Chinese robots and key components from entering the US market.
• Commercial ROI risk: Customers still need to demonstrate robot output, maintenance costs, intervention rates, and payback periods in real-world scenarios.
• Capacity misinterpretation risk: Factory construction and theoretical capacity do not equal actual shipments, yield, repeat purchases, or sustainable demand.
• Technology-path risk: Robot AI model approaches are still changing frequently, and insufficient real-world data could weaken model generalization.
• Safety-certification risk: As humanoid robots move from fenced demonstrations into shared industrial environments, safety validation and third-party certification could become key barriers.

What to watch

• Operating hours, intervention rates, maintenance costs, and incident rates in real deployments over the next several quarters.
• Repeat orders and robot-fleet expansion following initial pilots, rather than merely partnership announcements.
• Actual output, yield, and customer-use data from projects such as Tesla Optimus, XPeng IRON, and Boston Dynamics Atlas.
• Progress of the US GUARD Act and whether it expands to Chinese-manufactured components and manufacturing partnerships.
• Actual procurement, acceptance, and continued use after local governments, central state-owned enterprises, and state-owned enterprises submit application plans in China.
• Order realization in actuators, LiDAR, sensors, reducers, ball screws, and batteries across the humanoid robot supply chain.
• Progress by Nvidia, Mistral AI, Apptronik, and others in robot AI, safety architectures, and training-data feedback loops.