Unitree's Listing Enthusiasm Highlights Humanoid Robot Commercialization, Data Competition, and Leader Differentiation

Overview

Using Unitree's post-listing market performance, operating data, and use of proceeds as an observation window, the report analyzes the commercialization trajectory of China's humanoid robot industry. Morgan Stanley's core conclusion is that, beyond surging capital-market enthusiasm, the industry is undergoing changes in three areas: application mix, competitive factors, and differentiation among companies. However, breakthroughs in general-purpose capabilities still depend on whether data scale can genuinely translate into generalization capabilities.

Core views

Unitree's share price rose approximately 460% on its first trading day, reaching a market capitalization of RMB342 billion, equivalent to approximately US$51 billion, reflecting strong capital-market enthusiasm for humanoid robots. Based on FactSet's projected revenue of RMB2.9 billion in 2026 and RMB4.9 billion in 2027, its price-to-sales ratios reached 116x and 69x, respectively, the highest among comparable companies. Retail investor participation was particularly active, with the public offering tranche more than 8,000 times oversubscribed, while free float accounted for only 7.4% of total shares outstanding. The report believes Unitree's strong brand recognition from appearing on CCTV's Spring Festival Gala in both 2025 and 2026 also reinforced market attention. Operationally, Unitree's revenue growth is slowing: revenue grew 49% year over year in 1H26, below the 76% recorded for full-year 2025; 2Q26 revenue grew 39% year over year, also below 1Q26's 68%. Meanwhile, the R&D expense ratio increased from 9% in 2025 to 12% in 1H26, while 1Q net profit was also weighed down by elevated selling and promotional expenses. Rather than interpreting these changes solely as a growth issue for an individual company, the report links them to shifts in the industry's application mix. The report believes the humanoid robot industry's focus is moving from education, R&D, entertainment, and data collection toward real-world commercial scenarios. In 1H26, industrial and commercial uses had already risen to approximately 65% of Chinese humanoid robot shipments; by contrast, 76% of Unitree's humanoid robot revenue in 2025 still came from education and R&D. Unitree's changing growth rate therefore also reveals a divergence between its existing business mix and the direction of incremental industry demand. Morgan Stanley expects the share of industrial and commercial applications to increase further in 2H26 and beyond. The competitive focus is also shifting from hardware demonstrations and prototype capabilities toward models and training data. Unitree plans to use 48% of its approximately RMB6 billion in IPO proceeds for “robot brain” R&D, indicating that algorithms, models, and data have become core investment priorities. The report identifies data as the primary bottleneck: according to industry discussions, the robotics sector could experience a capability leap similar to the “GPT-3 moment” when training data reaches approximately 10 million hours. Leading companies have currently accumulated hundreds of thousands to several million hours of training data, and the approximately 10-million-hour level may be reached in 2027–2028. However, the report explicitly states that it remains uncertain whether this volume of data can produce the expected breakthrough so quickly. What genuinely needs to be validated is the robotics Scaling Law—whether expanding data can translate into stronger generalization capabilities and better real-world task performance. Industry concentration is already high. According to SAG data, the five largest Chinese humanoid robot integrators accounted for 87% of global shipments in 1H26. As the industry moves from demonstrations and prototypes toward commercialization, scale, data accumulation, and financing capabilities will become more important than before. Morgan Stanley expects the gap between leading companies and other participants to widen. As one of the world's largest humanoid robot companies, Unitree's financial performance and business updates can therefore provide an industry reference for covered companies with humanoid robot exposure, including Leaderdrive, Hengli Hydraulic, and Shuanghuan. On the first day of the World Robot Conference on August 19, 2026, the report observed that public interest remained strong and that the number of participating companies, product categories, and range of applications were all expanding. However, robots still lagged human-level capabilities by a significant margin. Among the rapidly growing applications, logistics sorting stood out as one of the fastest-improving scenarios, with its efficiency gradually approaching commercial requirements. This case supports the conclusion that real-world industry adoption is accelerating, while also indicating that commercialization will first emerge in scenarios with clearly defined task boundaries and readily measurable efficiency.

Analysis framework

The report first uses Unitree's first-day trading performance, valuation, subscription enthusiasm, and free float to assess capital-market attention. It then analyzes the company's revenue growth, expense ratios, and application mix, comparing them with the shipment uses of China's humanoid robot industry. Finally, through Unitree's use of IPO proceeds, training-data scale discussed in industry interviews, industry shipment concentration, and on-site observations from the World Robot Conference, the report sequentially assesses the competitive focus, conditions for capability breakthroughs, differentiation among industry leaders, and progress in commercial applications.

Methodology notes

• Valuation MethodologyPS valuation

  Comparable-company price-to-sales ratios

  The report divides Unitree's market capitalization by FactSet's projected revenue to derive price-to-sales ratios of 116x for 2026 and 69x for 2027, then compares them with peers to characterize the valuation the market is willing to pay for its future revenue growth.

• Event-Driven Strategies and Behavioral FinanceEvent-driven analysis

  Industry read-across from first-day IPO performance

  The report treats Unitree's first-day gain, oversubscription, and free float as event signals to gauge capital-market attention toward the humanoid robot theme and further analyzes the event's reference value for related industry companies.

• Competitive and Strategic FrameworkProduct life cycle

  Transition from demonstrations and R&D to commercial applications

  Based on shipment uses and observations from the World Robot Conference, the report concludes that the industry is moving from education, R&D, demonstrations, and prototypes toward industrial and commercial deployment, and uses this assessment to discuss changes in the factors driving corporate competition.

• Industry/Sector Analysis FrameworkIndustry Concentration Analysis

  Shipment share of leading integrators

  The report measures industry concentration using data showing that the five largest Chinese humanoid robot integrators accounted for 87% of global shipments and concludes that scale, data, and capital will further drive differentiation among industry leaders.

• (Out-of-Vocabulary Method)

  Observation of the robotics Scaling Law

  The report tracks whether the scale of training data can translate into generalization capabilities and real-world robot performance. It treats approximately 10 million hours of data as an important potential validation point for a capability leap rather than an established conclusion.

Asset mapping & comparison

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

• Unitree (Not Covered)

  The report uses its listing performance, financial data, application mix, and R&D investment as the primary window for observing the global humanoid robot industry.

  Strengths

  It is one of the world's largest humanoid robot participants, has strong public brand recognition, and its listing proceeds provide financial support for model and data R&D.

  Weaknesses

  Revenue growth slowed to 49% in 1H26, while 76% of its humanoid robot revenue in 2025 came from education and R&D, creating a structural mismatch with the industry's shift toward commercial applications.

  Comparison

  Its price-to-sales ratios of 116x and 69x based on 2026 and 2027 revenue forecasts are the highest among comparable companies.

  Risks

  Training data remains a bottleneck, and whether data scale can translate into generalization capabilities has not yet been validated.

• Leaderdrive（688017.SS）

  Morgan Stanley identifies it as a covered company with humanoid robot exposure that can use Unitree's industry developments as a reference.

• Hengli Hydraulic（601100.SS）

  Morgan Stanley identifies it as a covered company with humanoid robot exposure that can use Unitree's industry developments as a reference.

• Shuanghuan（002472.SZ）

  Morgan Stanley identifies it as a covered company with humanoid robot exposure that can use Unitree's industry developments as a reference.

Key data

• Unitree's first-day listing gain+460%Performance on the first trading day
• Unitree's first-day market capitalizationRMB342 billion/US$51 billionMarket capitalization reached after the first trading day
• 2026/2027 revenue forecastsRMB2.9 billion/4.9 billionFactSet revenue forecasts
• 2026/2027 price-to-sales ratios116x/69xDescribed by the report as the highest among comparable companies
• Retail oversubscriptionMore than 8,000xIndicates exceptionally strong retail investor participation
• Free float as a percentage of total shares outstanding7.4%Wind data
• Unitree's 1H26 revenue growth+49% YoYBelow +76% YoY in 2025
• Unitree's quarterly revenue growth2Q26 +39% YoY, 1Q26 +68% YoYGrowth slowed in 2Q relative to 1Q
• R&D expense ratio12% in 1H26, 9% in 2025R&D investment as a percentage of revenue increased
• Share of industry shipments for industrial and commercial usesApproximately 65%Shipment mix of Chinese humanoid robots in 1H26
• Unitree's share of revenue from education and R&D76%Humanoid robot revenue mix in 2025
• Share of IPO proceeds allocated to “robot brain” R&D48%Based on approximately RMB6 billion in IPO proceeds
• Training data required for a potential capability leapApproximately 10 million hoursPotential threshold for a “GPT-3 moment” mentioned in industry discussions
• Potential timing for reaching 10 million hours2027/2028Potential timing proposed by the report, although whether it will produce a capability breakthrough remains uncertain
• Global shipment share of the five largest Chinese integrators87%SAG data for 1H26

Impact & implications

The report believes the significance of Unitree's IPO extends beyond demonstrating market enthusiasm: it also reveals that the humanoid robot industry is shifting from R&D demonstrations toward commercial use. Companies' future relative positions will increasingly depend on models, training data, scale, and access to capital; leading participants with these resources may further expand their advantages. Clearly bounded tasks such as logistics sorting are approaching commercially viable efficiency, but broader capability breakthroughs still require validation of the data Scaling Law and real-world robot performance.

Risks

• Training data remains a key industry bottleneck, and it is uncertain whether approximately 10 million hours of data can produce a capability leap similar to the “GPT-3 moment.”
• Expanding data scale may not necessarily translate into robot generalization capabilities; the Scaling Law still requires validation through real-world performance.
• Despite the continued increase in products and applications, robots still lag human-level capabilities by a significant margin.

What to watch

• Track whether the share of industrial and commercial uses continues to increase in 2H26 and beyond.
• Monitor when leading companies' training data approaches approximately 10 million hours and whether data scale can translate into generalization capabilities.
• Continue validating robot performance in real-world tasks rather than observing only data volumes or demonstration results.
• Watch whether logistics sorting efficiency genuinely reaches a sustainable level of commercial viability.
• Track whether scale, data accumulation, and financing capabilities further widen the gap between leading companies and other participants.