Humanoid robots nearing a turning point: a thousand ships set sail, moving from "ten thousand deliveries" to the historic window of "million-scale production"
```
The humanoid robot industry is currently at a historic crossroads featuring a “fourfold resonance”: the start of mass production by overseas leaders, China's supply chain first-mover advantage, accelerated AI brain training, and continuous iteration of hardware solutions.
Tesla Optimus Gen-3 and subsequent models will start formal mass production in 2026, marking a shift from capital narratives to industrial execution in this field. It is expected that global mass production scale will leap from thousands to tens of thousands of units in 2026. The global division of labor—“American brain + Chinese body”—is gradually taking shape. Led by key players such as Sanhua Intelligent Controls and Tuopu Group, Chinese suppliers have successfully embedded themselves in the Tier-1 system of global core OEMs through pre-deployment of production capacity in Mexico and Thailand.
A new industrial storm is at its critical point. How should one understand and position the value of the industry chain?
I. What happened? Humanoid robots are approaching an inflection point
The humanoid robot industry is currently at a historic window featuring a fourfold resonance: overseas leaders launching mass production, China's supply chain first-mover advantage, accelerated AI brain training, and continuous iteration of hardware solutions.
Tesla Optimus Gen-3 and subsequent models will start formal mass production in 2026, marking a shift from capital narratives to industrial execution. It is expected that global mass production scale in 2026 will leap from thousands to tens of thousands of units. The global division of labor—“American brain + Chinese body”—is gradually taking shape. Led by key players such as Sanhua Intelligent Controls and Tuopu Group, Chinese suppliers have successfully embedded themselves in the Tier-1 system of global core OEMs through pre-deployment of production capacity in Mexico and Thailand.
In the short term, the industry sees three catalysts waiting to ferment—
① Tesla Optimus’s “Super Factory” momentum
With the completion of Tesla's Nevada factory’s dedicated humanoid robot production line, market attention has shifted from “can it walk” to “single-shift production capacity.” The certainty of the Optimus supply chain is the core logic of this year's overseas mainline.
② “Seed players” landing in subdivided scenarios
Humanoid robots have already entered three core scenarios first—automobile assembly, logistics sorting, and power inspection. It is expected that in the second half of 2026, the first electric cars manufactured with robotic participation will officially roll off the line, forming an “robots make robots” industrial closed loop.
③ “A thousand ships” competing among domestic body manufacturers
Domestic body manufacturers, represented by Unitree, UBTECH, and Fourier, are capturing the education, scientific research, and basic service markets with high cost performance and iterative speed, providing a strong supplement to the Tesla solution.
2025 is the “First Year of Mass Production” for humanoid robots, with global shipments skyrocketing from about 3,000 units in 2024 to 18,000 units (IDC estimate), a year-on-year increase of about 508%. In 2026, the industry will shift from “ten thousand deliveries” to “million mass production”—Tesla Optimus Gen3 will begin production in summer 2026, Fremont factory is being transformed into a super base with annual production of 1 million robots; Unitree aims to ship 10-20 thousand units, UBTECH has significantly increased its Walker S series shipment target to 5,000 units, and Zhiyuan Robotics’ 10,000th Exped A3 rolled off the line at the end of March 2026.
2025 is a key turning point for the humanoid robot industry to go from proof of concept to large-scale mass production. According to IDC's “Global Humanoid Robot Market Analysis,” global humanoid robot shipments will be about 18,000 units in 2025, a year-on-year increase of around 508%, with Chinese manufacturers leading the global market.
Into 2026, institutions have further raised growth expectations for humanoid robots. UBS forecasts global demand will reach 30,000 units in 2026 and up to 40,000 units in an optimistic scenario. Morgan Stanley raised its forecast for China's 2026 humanoid robot sales from 14,000 units to 28,000 units. Bank of America is more aggressive, predicting annual shipments will rise from 90,000 units in 2026 to 1.2 million units in 2030, with a compound annual growth rate of 86%, faster than the early electric vehicle market.
Looking ahead to 2030 and beyond, Counterpoint predicts global humanoid robot shipments may reach 256,000 units, with a compound growth rate of 69.7%; Omdia forecasts that in 2035 global shipments may exceed 2.6 million units, about 150 times the shipments in 2025. Morgan Stanley expects China's sales will reach 262,000 units in 2030 and 2.6 million units in 2035.
Unlike the overseas prudence, Chinese humanoid robot firms are seeing explosive growth in 2025.
China has formed a significant scale advantage in the humanoid robot industry. From the perspective of company shipment shares, Chinese firms shine: Unitree shipped over 5,500 units in 2025, accounting for 32.4% of the global market; Zhiyuan shipped over 4,000 units, accounting for 23.5%; Leju, Accelerated Evolution, and Songyan Power shipped around 1,000 units each, each accounting for 5.9%; UBTECH shipped about 600 units, accounting for 3.5%. These six companies together hold 74.1% of global shipment share.
Different from the large-scale mass production of Chinese companies, overseas leading companies are still focusing on technological breakthroughs and ecological synergy, and have not achieved scaled mass production. Tesla's 2025 Optimus trial production scale is less than 1,000 units, and Figure AI, valued at nearly RMB 280 billion in 2025, shipped only about 150 units for the entire year. This contrast highlights China’s first-mover advantage in humanoid robot hardware manufacturing and commercialization.
TrendForce predicts China's humanoid robot market output will rise by 94% year-on-year in 2026, with Unitree and Zhiyuan together occupying about 80% of domestic shipment quantity, forming a concentrated leading production structure.
II. Why is it important? An unignorable major trend
The industrialization progress of Tesla Optimus is the biggest catalyst for the global humanoid robot industry in 2026. In April 2026, Musk made it clear during an earnings call: Optimus Gen3 will begin production in summer, and move to large-scale mass production in 2027. Fremont factory’s Model S/X lines are being rebuilt into a super base producing 1 million robots annually.
This decision has very profound strategic significance. Tesla is discontinuing its flagship Model S and Model X to make way for Optimus, and Musk bluntly said “Ultimately, no one will remember that Tesla ever made cars.” Musk's trillion-dollar compensation plan is directly tied to the goal of “producing 1 million robots,” making Optimus not a side business but Tesla’s biggest bet as a trillion-dollar company.
On product iteration, Optimus Gen3 will see many core upgrades: a brand new dexterous hand design, increasing the freedom degree from Gen1’s 6 to about 25 per dexterous hand (including forearm); adopting a tendon-rope scheme, placing the drive module at the wrist for greater flexibility and lower terminal weight; substantial increase of planetary roller screws concentrated in high-load joints like legs and waist.
Musk’s attitude towards Chinese competitors is both wary and confident—he frankly said “Tesla humanoid robot's biggest competitors are definitely in China,” citing Chinese companies as “very good at scale and manufacturing, and strong in AI capability”; yet also stressed Optimus is “stronger than any known Chinese humanoid robots, especially in AI, hand design, and electromechanical dexterity.” This keenly reveals the essence of current global competition: China leads on hardware manufacturing and scaled delivery, but the US retains advantages in AI algorithm and high-precision electromechanical design.
Domestic humanoid robots’ application abilities in single scenarios are rapidly improving. Though efficiency still has room for improvement, they already have preliminary replacement capabilities, with commercial demonstration, guided touring, and factory handling already validated.
① Industrial handling is one of the most mature current scenarios. UBTECH has deep layout in industrial field, with handling, sorting, and data collection tasks gradually maturing. In 2025, the company’s humanoid robot order reached nearly 1.4 billion RMB, recognized by leading customers in automotive and 3C fields. The company’s new full-chain solution can reduce data labeling costs by 99%.
② In commercial services, Unitree has frequently gone viral—showcasing complex and difficult movements at large concerts, and complex operations at the 2026 Spring Festival Gala (golden cudgel, nunchucks, aerial rotations, etc.), demonstrating precise and stable motion control. Xiaopeng Robot's commercialization strategy is clear, expected to be first applied in “three guides” (tour guide, shopping guide, patrol guide), with a 2030 goal of 1 million units.
③ In industrial inspection scenarios, humanoid robots can be used in mine inspection, mining and other hazardous operations—playing vital roles in mining, chemical, and power fields. For power inspection, humanoid robots’ architecture matches the adaptability required for complex environments, with dexterous hands enabling fine operations of power facilities.
④ Scientific research and education scenarios are also rapidly penetrating. UBTECH and Beijing Robotics Innovation Center jointly launched a full-size research and education humanoid robot platform “Tiangong Walker,” starting at 299,000 RMB—the industry’s first research-grade humanoid robot below 300,000 RMB.
The global humanoid robot competitive landscape is now forming a tripartite competition among China, the United States, and Europe.
China: The global leader in hardware manufacturing and scaled delivery. Leveraging a complete supply chain, strong manufacturing capabilities, and policy support, China has established absolute lead in whole-machine shipments. By the end of 2025, China had over 140 whole-machine firms, with the global top 6 shipment companies all from China. Chinese firms’ core strength: rapid technical response capability, stable localized supply of raw materials, and ability to quickly reduce costs through scale.
United States: Leader in AI algorithms and electromechanical design. The US’s strength centers on the “brain” layer—AI large model training, reinforcement learning algorithms, and high-precision electromechanical design. Tesla Optimus remains ahead in AI, dexterous hand design, and mechanical dexterity; Figure AI, Boston Dynamics, etc., have deep expertise in motion control and AI algorithms.
Europe: Supplemental force in industrial scenarios and basic research. Europe has deep roots in industrial automation, precision machinery, and robotic basic research, but relatively slow commercial progress in whole-machine humanoid robots.
Notably, both competition models are rapidly merging—US companies use Chinese supply chains for hardware production (Tesla’s Chinese suppliers comprise 60%-70%), while Chinese firms are catching up on AI algorithms and software capacity, blurring technological boundaries.
III. Next focus? Panoramic supply chain scan
The humanoid robot industry is crossing from “technical verification” to “commercial mass production,” divided into three stages:
Stage One (2024-2025): Technical breakthrough and mass production start. 2024-2025 marks the key point moving from labs to production lines. Unitree shipped over 5,500 units in 2025, Tesla began Fremont Optimus line conversion, Zhiyuan completed offlining 10,000 units. We’re currently accelerating towards Stage Two.
Stage Two (2026-2028): Scale mass production and order verification. This is the core current stage. Tesla Optimus Gen3 production starts in summer 2026, Unitree aims for 10-20 thousand units, UBTECH raises target to 5,000 units. The key issue shifts from “who can manufacture” to “who can prove value in real scenarios”—order realization, capacity ramp, and gross margin improvement are the three main metrics.
Stage Three (2028-2030): Global competition and scenario diversification. As technology matures and costs fall, humanoid robots will move rapidly from B-end to C-end, into broader applications, and global competition will become clearer.
Tesla’s Optimus supply chain offers the most complete model for understanding humanoid robot supply chain structure. Starting June 2025, Tesla began Optimus mass production factory audits, finally selecting seven companies as core suppliers: Lens Technology, Sanhua Intelligent Controls, Tuopu Group, Changying Precision, Wolong Electric Drives, Wuzhou New Spring, and Hengli Hydraulic. Chinese suppliers account for about 60%-70% of Optimus core component value; Gen3 unit cost is around $20,000, with core component value over 70%. Chinese supply chains are not just key to Tesla's cost reduction and efficiency goals, they've formed a complete cycle from “core components—drive systems—actuators—whole machine assembly.”
Core conclusions—
Conclusion 1: 2026 is the pivotal watershed from “concept validation” to “order validation.” The 508% shipment surge in 2025 proved the explosive “zero-to-one” strength; in 2026, leading manufacturers’ ten-thousand level delivery goals, Tesla’s million-unit line transformation, and UBTECH’s increased shipment targets mark the move from “ten-thousand deliveries” to “million-unit mass production.”
Conclusion 2: Chinese firms have established global shipment leadership, but AI brain and precision design remain weaknesses. In 2025, Chinese shipments account for over 86% of the global total, but Tesla remains ahead in AI and dexterous hand design. The core competition is shifting from “hardware scale race” to “integrated software-hardware competition.”
Conclusion 3: Supply chain value is upgrading from “hardware manufacturing” to “core components + system integration.” Planetary roller screws, dexterous hands, electronic skin, high-precision reducers—these core components have the highest technical barriers and value. In Optimus Gen3, core component value is over 70%, with Chinese suppliers at 60%-70%; firms mastering these will gain higher valuation premiums.
Conclusion 4: Two clear investment mainlines in 2026. Overseas, focus on Tesla Optimus industrial progression and the seven core audited suppliers. Domestically, pay attention to leaders in body plus application with strong commercialization and order realization capability.
Risk warning and disclaimerThe market has risks, and investment should be cautious. This article does not constitute personal investment advice, nor does it consider individual users’ specific investment objectives, financial situation, or needs. Users should decide whether any opinions, views, or conclusions in this article suit their particular circumstances. Investments made accordingly are at the user's own risk. ```