The first round of the "elimination match" has begun! Morgan Stanley announces: humanoid robots have entered the mass production phase of "knife fighting," and PPT players are being eliminated

Morgan Stanley warns that China's humanoid robot industry has moved from the PPT presentation stage into a real production competition phase.

According to information from the Chasing Wind Trading Desk, Morgan Stanley's research team conducted intensive visits to several complete machine manufacturers (Fourier, Kepler, MagicBot) and core component suppliers (Greens Harmonics, Hengli Hydraulic, Shuanglin Co., Ltd., Zhenyu XieNeng, Flat, Wolong Electric Drive, Siling Technology) from January 26 to 28. Analysts observed a key change: the gap between leaders and laggards is rapidly widening, and the first round of industry reshuffling is imminent.

Morgan Stanley analysts believe that core component suppliers with technological strength and mass production capabilities will dominate this wave of industrialization. As the industry's shipment volume is expected to achieve several times growth by 2026, suppliers with first-mover advantages will gain significant market share and pricing power.

Shipment volume will achieve several times leap in 2026

Morgan Stanley's research shows that all complete machine manufacturers hold optimistic expectations for shipment volume in 2026. A leading domestic complete machine manufacturer has already exceeded 5,000 units in shipment volume in 2025, with expectations for several times growth in 2026. Other manufacturers' specific targets include: Fourier aims for 2,000 units (400-500 units in 2025); MagicBot aims for over 1,000 units; Kepler aims for 300 units (70-80 units in 2025).

Regarding component suppliers, Hengli Hydraulic and Siling Technology are preparing production capacity based on the production plans of North American complete machine manufacturers—this client plans to achieve a capacity of 1,000 units per week by July 2026, further increasing to 2,500 units per week by the end of the year. Greens Harmonics expects its embodied robot business to achieve exponential growth, potentially matching its industrial robot shipment volume.

The key support behind these numbers comes from government-backed projects. Early shipment volumes still mainly rely on R&D and government-supported project procurement. Most complete machine manufacturers visited during the research have shipment records to data collection centers or have received government support.

Task capability differentiation: Who is really making products

Morgan Stanley observed a decisive difference during its field investigations: there is a clear differentiation in task capability and execution efficiency/speed among complete machine manufacturers. As the industry bids farewell to the video demonstration phase in 2025, the increased customer feedback and actual shipments in 2026 will further reveal this gap.

This differentiation is not only reflected in shipment volumes but, more critically, in who can complete the closed loop from actual deployment to iterative improvement—including model optimization, task capability enhancement, and cost control. Although unit shipment volumes are rapidly increasing, the improvement speed of robotic operational capabilities remains slow, constrained by models, data, and computing power.

Morgan Stanley assesses that lagging startups will find it harder to catch up during the industry's acceleration phase, and the first round of reshuffling may happen soon. Additionally, component suppliers reveal that companies from other industries, such as XPeng and Xiaomi, are also accelerating their robot development.

Core component suppliers expand leading advantages

The component supplier market also shows a clear advantage for leaders. Suppliers with stronger technological strength and mass production capabilities (better quality, consistency, yield, cost, and overseas supply readiness) are expanding their market leadership, gaining higher penetration rates among leading complete machine manufacturers Suppliers are shifting from single-component competition to module-level product supply, including actuators, modules, motors, bearings, and even self-produced equipment. This strategy aims to help customers reduce integration complexity, improve consistency and quality control, thereby obtaining higher value and profit margins, and avoiding price competition on single components.

Although the industry is still in its early stages, overseas expansion has become a key focus.

Leading domestic manufacturers contributed a "single-digit" percentage to overseas sales last year and only configured overseas sales teams in the second half of 2025, with stronger growth expected in international business this year. Demand from North America tends to focus on R&D and data collection.

Application Scenario Validation: Gradual Progress, No Universal Breakthrough Yet

Manufacturers are seeking repeatable and scalable application cases in multiple scenario pilots, including industrial, retail, medical, logistics, guidance/display, etc.

Morgan Stanley continues to anticipate a gradual development trajectory: a slow, scene-by-scene advancement driven by a human-machine collaborative data flywheel, rather than a rapid, unified universal operational breakthrough.

Leading domestic manufacturers expect that one-third of shipments in 2026 will come from entertainment/commercial services, one-third from industrial/data collection, and the remainder from R&D.

Fourie expects that 60% of revenue in 2026 will still come from R&D/education/data collection, 10% from entertainment, and the rest from industrial and elderly care. Kepler currently has application cases in logistics and industrial unloading (in the concept validation stage), aiming to enter commercial services (cinemas/coffeeshops) by 2026. MagicBot focuses on industrial unloading/inspection, guidance/display, unmanned retail, and pharmacies.

These application validations indicate that the commercialization of humanoid robots still requires time, but the industrialization process is irreversible, and companies with real technological strength and mass production capabilities will win in this race