Humanoid robotics: Humanoid capacity is scaling, but deployment evidence and investor sentiment still lag long-term demand expectations

Overview

This UBS thematic update examines whether expanding humanoid-robot capacity is translating into real adoption, using China and Europe C-suite surveys alongside industry developments. The report sees durable long-term demand, but emphasizes that robot intelligence, commercial conversion and proof of large-scale production remain the critical near-term constraints.

Core views

The prototype-to-production transition is progressing globally, although the "robot brain" remains the principal bottleneck. UBS cites UBTECH bringing its Liuzhou factory with planned annual capacity above 10,000 units into production, XPeng starting an automated humanoid assembly line, and reducer makers Leader Drive and Laifu planning aggressive capacity expansion. Internationally, OpenAI has re-entered humanoids, Figure AI has obtained access to up to 100,000 NVIDIA GPUs, and Toyota has unveiled the wheeled humanoid ELEY. Media reports cited by UBS indicate Toyota plans to deploy 400,000 robots, including humanoid and non-humanoid systems, across global manufacturing and affiliated suppliers from 2028. UBS’s China channel checks also indicate that Tesla Optimus suppliers are gradually ramping production. Regulatory scrutiny has become a near-term headwind but is viewed by UBS as potentially constructive for the industry’s longer-term development. Potentially tighter IPO scrutiny for Chinese humanoid companies and concerns over start-up revenue quality have hurt sentiment; UBS argues that higher standards could improve industry quality, accelerate consolidation and support more sustainable growth. In the US, FCC restrictions announced on 28 July on certain foreign-made advanced robotic systems could restrict Chinese robot OEM access to the US market unless they demonstrate credible localization and supply-chain compliance. UBS expects limited near-term effect on component suppliers because the restrictions principally target finished systems. UBS Evidence Lab surveys point to stronger adoption interest in China than Europe. In China’s June 2026 survey, 64% of respondents would consider humanoids to replace labor and 45% were already testing them on production lines, compared with 51% and 38%, respectively, in Europe’s May 2026 survey. UBS estimates that China is roughly one to two years ahead in commercialization and deployment. The gap appears mainly to reflect willingness rather than execution, as testing equals roughly 70% of adoption intent in both regions. Technology and consumer-goods manufacturing, including autos, rank above average for adoption intent in both regions, whereas several services, retail, construction and staples categories rank below average. The China survey also highlights a disconnect between long-term conviction and near-term conversion. Over five survey waves since May 2024, production-line testing fell to 45% in June 2026 from 57% in May 2024 and has not expanded over the past two years, while adoption intent has remained above 50% in every wave and reached 64% in June 2026. Cost remains a major barrier among companies not considering humanoids, and UBS notes that new production capacity is running well ahead of current sales. It therefore identifies repeat orders and capacity utilization as key near-term indicators. Longer-term signals remain favorable in UBS’s view. Ninety-three percent of China survey respondents expect humanoid commercialization within five years. Willingness to pay improved, with 36% willing to pay more than Rmb400,000 per unit in June 2026 versus 24% in May 2024, while safety concerns have eased. UBS expects cost reductions from scaling capacity to narrow the gap between market prices and customer willingness to pay. It cautions that the June survey sample was relatively small and therefore focuses on multi-wave trends rather than wave-on-wave movements. Use-case expectations reinforce the importance of AI and dexterity. Chinese companies show relatively greater interest in sorting, reflecting manufacturing scale and automation demand. In both China and Europe, assembly is cited at least as often as loading and unloading, implying that users expect robots to move beyond basic material handling. UBS concludes that this progression depends on improved dexterity and AI, consistent with its view that intelligence remains the central constraint on broader adoption. UBS previously raised its 2026-30 global humanoid demand forecasts by more than 70% annually, reflecting faster uptake in preparatory and lower-complexity applications. It also forecasts 2025-30 China demand CAGRs of 10-60% across traditional robot categories, with AI serving as a catalyst for wider robotics adoption. However, the report says sector sentiment is unlikely to improve meaningfully until Tesla Optimus V3 mass production, AI-model generalisation and hardware advances become clearer. Within the theme, UBS prefers sensors, reducers and roller screws because they can serve multiple robot-OEM categories, offer attractive market sizes and have high technological barriers; its China top picks are OPT, Kedali and Hengli Hydraulic.

Analysis framework

UBS combines multi-wave China and Europe C-suite survey results with industry channel checks, company developments, supply-chain mapping and market-performance observations. It compares adoption willingness, production-line testing, use cases, willingness to pay and barriers across regions, then links these findings to capacity expansion, technical constraints and component exposure.

Methodology notes

• Industry AnalysisSupply-demand framework

  Humanoid supply-demand and commercialization analysis

  UBS compares rapidly expanding manufacturing capacity with current sales, adoption intent, testing activity, willingness to pay and expected use cases to assess the path from prototypes to broader deployment.

• Other

  China and Europe C-suite business surveys

  UBS Evidence Lab surveys senior executives to measure labor-replacement willingness, production-line testing, purchase barriers, expected applications and commercialization expectations.

Asset mapping & comparison

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

• OPT Machine Vision (688686.SH)

  China top pick and sensor/vision component exposure within the humanoid supply chain.

  Strengths

  UBS prefers sensors because they can serve multiple robot OEM categories and have high technological barriers.

  Comparison

  Included among UBS’s preferred upstream component exposures alongside Kedali and Hengli Hydraulic.

  Risks

  Worse-than-expected downstream capacity expansion, gross-margin pressure from product mix or in-house product-ratio improvement, and higher SG&A from new-business development.

• Shenzhen Kedali Industry (002850.SZ)

  China top pick with component exposure relevant to robotics and broader industrial supply chains.

  Strengths

  UBS identifies Kedali as one of its China top picks in the humanoid theme.

  Comparison

  Included among UBS’s preferred upstream component exposures alongside OPT and Hengli Hydraulic.

  Risks

  Slower capacity building at the company or customers, customer share losses, product-quality issues, weaker EV penetration, pricing pressure, competition, raw-material inflation and trade friction.

• Jiangsu Hengli Hydraulic (601100.SH)

  China top pick and upstream component supplier exposed to robotics-related demand.

  Strengths

  UBS identifies Hengli Hydraulic as one of its China top picks in the humanoid theme.

  Comparison

  Included among UBS’s preferred upstream component exposures alongside OPT and Kedali.

  Risks

  Weaker domestic excavator sales, weaker pumps-and-valves demand, failure to enter overseas-brand supply chains, weaker non-standard-product demand and trade friction affecting overseas expansion.

Key data

• China adoption intent64%Respondents willing to consider humanoid robots to replace labor in the June 2026 China survey.
• China production-line testing45%June 2026, compared with 57% in May 2024.
• Europe adoption intent and testing51% and 38%May 2026 survey readings, versus 64% and 45% in China.
• Expected commercialization within five years93%Share of China respondents expecting humanoid commercialization within five years.
• Willingness to pay above Rmb400,000 per unit36%June 2026 versus 24% in May 2024.
• Global humanoid demand forecast revisionMore than 70% annuallyUBS upgrade to its 2026-30 global demand forecasts in its prior APAC Focus report.
• China traditional-robot demand CAGR10-60%UBS forecast for 2025-30 across traditional robot categories.

Impact & implications

UBS sees the industry shifting toward scaled production, but adoption must catch up with capacity. The report’s preferred exposure is through technically demanding, cross-OEM component suppliers rather than a broad theme approach, while clearer evidence on repeat orders, utilization, AI capability and Optimus production is needed to improve sentiment.

Risks

• Robot intelligence, dexterity and AI-model generalisation may advance more slowly than needed for broader adoption.
• New humanoid production capacity is running ahead of current sales, and weak repeat orders or low utilization could delay commercialization.
• Tighter China IPO scrutiny and US restrictions on certain foreign-made advanced robotic systems could weigh on sentiment or limit Chinese OEM market access.
• China industrial demand could weaken if macro investment remains soft, policies change or competition causes market-share losses.

What to watch

• Repeat orders and capacity utilization at humanoid producers and suppliers.
• Clarity on Tesla Optimus V3 mass production and the timing of its ramp.
• Advances in AI-model generalisation, robot intelligence and hardware capability.
• Progress in assembly use cases, where dexterity is essential.
• Upcoming industry events, company results and Tesla’s planned end-2026 Optimus Gen3 mass-production start.