Who Leads the Humanoid Robot Race? A Multi‑Dimensional Scoring Model Ranks the Top 30 Companies

Model Design

The article constructs a multi‑dimensional weighted scoring model to evaluate humanoid robot companies across five key dimensions. The dimensions and their weights are derived from industry analyses (IDC, Omdia, Yole Group, Morgan Stanley, etc.).

Dimensions and Weights

D1 – Motion Control & Hardware (25%) : extreme motion verification, degrees of freedom, joint precision, in‑house hardware development.

D2 – AI & Embodied Intelligence (20%) : maturity of VLA models, multi‑task generalisation, depth of collaboration with top AI labs.

D3 – Mass Production & Delivery (25%) : annual shipment volume, production line scale, supply‑chain integration.

D4 – Commercial Deployment (20%) : annual order/revenue amount, number and quality of paying customers, contract renewals.

D5 – Capital Strength (10%) : valuation/market cap, total financing, strategic investors.

Scoring Anchors

Each dimension has concrete score brackets that can be verified against public data.

D3 – Mass Production

Score 9‑10: >3000 units shipped annually (e.g., ZhiYuan Robotics – 5168 units, Omdia).

Score 7‑8: 1000‑3000 units (e.g., Yushu Technology – ~1500 units, 2024).

Score 5‑6: 300‑1000 units with mature production line (e.g., UBTech – 500+ units, capacity 1000).

Score 4‑5: 100‑300 units, ramping up (e.g., Agility Robotics – ~100 units; Figure AI – estimated ~1000 units, not yet full production).

Score 3‑4: <100 units or no external sales (e.g., Tesla Optimus – several hundred units for internal use).

Score 2‑3: prototype or pilot stage (e.g., Boston Dynamics – announced mass production at CES 2026).

D4 – Commercial Deployment

Score 9‑10: >5 billion CNY annual orders, multi‑industry deployment (e.g., UBTech – ~14 billion CNY).

Score 7‑8: >5 billion CNY revenue, multiple paid channels (e.g., Yushu Technology – >10 billion CNY, B‑end + research + consumer).

Score 6‑7: several real paying customers with long‑term contracts (e.g., Agility Robotics – Amazon, GXO, Spanx).

Score 4‑5: few POC or early deployments (e.g., Figure AI – BMW pilot, limited customers).

Score 3‑4: no external commercial revenue (e.g., Tesla Optimus, Boston Dynamics Atlas).

D1 – Motion Control

Score 10: world‑class extreme actions (back‑flips, parkour) with full electric drive.

Score 8‑9: high‑difficulty actions publicly demonstrated, full‑stack hardware self‑development.

Score 6‑7: stable walking and fine manipulation, no extreme action verification.

Score 5‑6: basic bipedal walking, manipulation capabilities under optimisation.

D2 – AI & Embodied Intelligence

Score 9‑10: self‑developed VLA model tightly integrated with Nvidia/DeepMind/OpenAI, multi‑task generalisation.

Score 7‑8: scenario‑oriented AI, self‑developed or heavily optimised action model, collaboration with leading AI institutions.

Score 5‑6: optimisation of open‑source models for specific scenarios, limited generalisation.

D5 – Capital Strength

Score 10: valuation > $100 billion.

Score 8‑9: valuation $10‑100 billion.

Score 7‑8: valuation $1‑10 billion.

Score 5‑6: valuation $0.1‑1 billion.

Weighted Formula

The overall score is the weighted sum of the five dimension scores:

OverallScore = 0.25·D1 + 0.20·D2 + 0.25·D3 + 0.20·D4 + 0.10·D5

Each company’s dimension scores are multiplied by the corresponding weight and summed to obtain a total score out of 10.

Top‑30 Ranking (selected highlights)

Using the formula, the top three companies are:

ZhiYuan Robotics (China) – Total 8.88: D1 8.5, D2 8.5, D3 10.0, D4 9.0, D5 7.5. 5168 units shipped (Omdia) and >10 billion CNY in orders.

Yushu Technology (China) – Total 8.23: D1 9.5, D2 7.0, D3 8.0★ (estimated), D4 8.5, D5 7.5. Electric‑driven humanoid capable of back‑flips, >10 billion CNY revenue, five consecutive years of profit.

UBTech (China) – Total 7.73: D1 7.5, D2 8.0, D3 6.0, D4 9.5, D5 8.5. Annual orders ~14 billion CNY, multi‑industry customers (BYD, Geely, Foxconn).

Other notable entries include Figure AI (USA) with a high valuation but low production score, Tesla Optimus with strong AI but minimal commercial deployment, and Boston Dynamics with perfect motion‑control but no 2025 shipments.

Key Insights

Chinese firms dominate the top‑four positions, reflecting a >2‑point lead over the highest‑ranked U.S. company. Their advantage stems from combined strengths in mass production (D3) and commercial orders (D4).

Valuation does not always translate to delivery capability; Figure AI’s $390 billion valuation (D5 = 10) contrasts with a modest production score (D3 = 4.5), while Tesla Optimus scores low on commercial deployment despite high hype.

Boston Dynamics illustrates the largest gap between technical excellence (high D1/D3) and market impact, scoring 6.08 overall due to zero commercial shipments in 2025.

Market Landscape

Omdia estimates global humanoid robot shipments in 2025 at ~13 000 units, with Chinese manufacturers accounting for >85%. GGII projects the market size at ¥63 billion in 2025, rising to ¥340 billion by 2030.

Model Limitations

Data availability is uneven; many Western firms lack public shipment or order figures, leading to estimated scores (marked ★).

The weight distribution reflects the current industry stage. A shift toward AI‑centric competition would increase D2’s weight, reshuffling rankings.

Static rankings become outdated quickly; valuation and production figures can change dramatically within months.

Figures