Unitree Technology Pre-IPO Survey: Full-Stack Cost Leadership, but 'Brain' Bottleneck Awaiting Resolution

Overview

This report records Nomura Securities' survey minutes of leading humanoid robot enterprise preparing for listing, Unitree Technology (Unitree), held on June 15, 2026. The report's core points indicate that Unitree Technology has already achieved significant profitability and global shipment leadership in 2025 due to cost advantages brought by full-stack self-research, and successfully passed STAR Market listing review. However, the core bottleneck for industry development is shifting from hardware motion control to general-purpose artificial intelligence "brain" (such as world models), and technical routes remain unified. Additionally, the report updates views on supply chain enterprise Tuopu Group, maintaining a neutral rating, questioning the sustainability of its contribution from humanoid robot business.

Core views

Unitree Technology demonstrates strong financial performance and market position. In 2025, the company's humanoid robot shipments exceeded 5,500 units, ranking first globally, revenue grew significantly by 335% year-over-year to 1.7 billion RMB, reaching net profit of 288 million RMB. Management revealed that the company's gross margin is approximately 60% and production-sales rate exceeds 90%. This performance is attributed to its "full-stack self-research" model, where core components such as motors, reducers, and LiDAR are all self-developed, with purchased components accounting for only 14-18% of total costs. Cash flow generated from mature quadruped robot business provides funding support for R&D and market expansion of these humanoid robots. Although hardware manufacturing capabilities lead the industry, the real constraints faced by the industry have shifted to the "brain" level, namely general-purpose AI technology. External industry surveys show that although Unitree leads globally in motion control hardware, the dominant model roadmap has not yet converged. Researchers are moving from Vision-Language-Action (VLA) models to world models, but there remain huge differences in the definition of "world" and scenario understanding. Since high-level paths and sub-paths are not determined, no clear timeline is given in the prospectus. Therefore, industry views believe hardware manufacturers may face increasingly intensified competition pressure from AI giants in the future. Unitree plans to use more than 50% of IPO raised funds for "brain" development to support long-term growth. In terms of application landing, demand in non-academic fields is scaling up rapidly. Although currently 74% of humanoid robot revenue still comes from universities and research institutions, power grid inspection and firefighting have become the fastest-growing industrial use cases. Management stated that the company positions itself as a general-purpose robot manufacturer, rather than a machine for specific scenarios, so expansion into other vertical fields mainly adopts a "demand pull" strategy, developing dedicated solutions together with partners when market demand arises.

Analysis framework

Nomura adopted an analysis method combining field research and cross-validation along the industrial chain. First, by directly interviewing Unitree Technology management, it obtained key financial data before IPO (revenue, profit, gross margin), technical architecture (full-stack self-research ratio), and strategic planning (fund usage, market strategy). Second, combined with external industry surveys, compared Unitree's leading position in motion control hardware with common industry bottlenecks in AI software levels, pointing out potential structural risks faced by hardware manufacturers in the AI era. Finally, by mapping Unitree's industry dynamics to listed company Tuopu Group, evaluates the beneficiary degree and risk of supply chain enterprises, thereby deriving a neutral rating judgment on Tuopu Group.

Methodology notes

• Industry / Sector Analysis FrameworkUpstream/Midstream/Downstream Supply Chain Transmission

  Value Distribution along Supply Chain and Shift in Technical Bottlenecks

  The report points out that the value focus of the robotics industry is shifting from upstream hardware (motion control) to midstream core algorithms (AI brain). This analysis helps investors understand why hardware-leading enterprises still face potential threats from AI giants, and why Unitree needs to invest heavily in software R&D.

• Company Fundamentals and Financial FrameworkProfitability Quality Analysis

  Contribution of Full-Stack Self-Research to Gross Margin

  By analyzing Unitree Technology's high gross margin of 60% and low proportion of purchased components (14-18%), the report reveals that the core source of its cost control is the vertically integrated full-stack self-research model, which is a key indicator for evaluating the sustainability of its profitability.

Asset mapping & comparison

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

• Unitree Technology (Unlisted)

  Survey Subject, Global Leader in Humanoid Robot Shipments, Preparing for STAR Market IPO

  Strengths

  Full-stack self-research brings 60% high gross margin, achieved profitability in 2025, motion control leads globally

  Weaknesses

  AI "brain" technical route undecided, non-academic field application proportion still low

  Comparison

  Compared with pure hardware manufacturers, possesses stronger cost control ability; Compared with AI giants, lags in general-purpose model R&D

  Risks

  AI technical route converges slower than expected, intensifying industry competition leads to hardware price wars

• Ningbo Tuopu Group (601689.SS)

  Relevant Enterprise in Robot Supply Chain, Neutral Rating Given by Report

  Weaknesses

  Sustainability and scale of humanoid robot business contribution questioned, core business growth expected to decelerate

  Comparison

  Compared with Unitree Technology's full-stack self-research, Tuopu as a component supplier is more affected by downstream mass production progress

  Risks

  Key customer shipments fall short of expectations, humanoid robot mass production progress slows down

Key data

• 2025 Humanoid Robot Shipments>5,500 unitsRanked First Globally
• 2025 RevenueRMB 1.7 BillionGrowth of 335% Year-over-Year
• 2025 Net ProfitRMB 288 MillionAchieved Profitability
• Gross Margin~60%Benefiting from Full-Stack Self-Research
• Purchased Components Cost Proportion14-18%Self-Developed Core Components
• Academic Field Revenue Proportion74%Non-Academic Demand Scaling Rapidly
• Tuopu Group Target PriceRMB 55Based on 34x 2026 Expected PE

Impact & implications

For Unitree Technology, its full-stack cost advantage has been validated by the market, but the future competitive high ground lies in AI model breakthroughs. If progress cannot be made in "brain" technology, its hardware advantage may be eroded by incumbents (market leaders) possessing strong AI capabilities. For supply chain enterprises like Tuopu Group, the report warns to view cautiously the long-term contribution scale of its humanoid robot business, believing its core business growth may slow due to lack of scale advantage. Overall, the robotics industry is at a critical stage transitioning from hardware-driven to software intelligence-driven.

Risks

• Humanoid robot mass production progress slower than expected
• AI model technical route fails to converge long-term, leading to wasted R&D resources
• Tuopu Group key customer shipments fall short of expectations
• Weak EV demand affects Tuopu core business

What to watch

• Effect of fund allocation for "brain" R&D after Unitree Technology IPO
• Continuous release of demand in non-academic fields such as power grid inspection and firefighting
• Convergence process of industry-dominant AI model roadmaps
• Growth trend of Tuopu Group core business in 2026