The physical AI opportunities identified by Tianfeng Securities’ machinery team center on which specific manual tasks products can cover. The corresponding worker pools provide demand context, while product form and development stage determine how these opportunities should be understood.
Which manual tasks correspond to the opportunities in apparel and packaging?
The entry points in apparel are handling fabric pieces and operating sewing equipment. Jack Technology’s products include AI sewing machines, AI cutting equipment, unmanned template machines and humanoid robots for apparel production. The robots focus on grasping, separating and positioning fabric pieces and operating template machines, extending automation to manual tasks before and after equipment operation.
Packaging encompasses case packing, case sealing, packing, labeling, palletizing, loading and unloading. Youngsun Intelligent’s product portfolio includes collaborative robots, handling robots and intelligent packaging lines; its wheeled humanoid robot has reached a finalized design and is intended for packaging and warehousing. A finalized design indicates development progress; this description in the materials does not specify delivery scale.
How do the product offerings of companies serving logistics differ?
Damon Technology’s products cover conveying, sorting, picking, depalletizing, palletizing and handling, corresponding to multiple processes in express delivery, warehousing and in-plant logistics. Its opportunity comes from equipment demand across different logistics tasks and cannot be understood solely in terms of a single handling role.
Hangcha Group and Anhui Heli both address forklift-driver roles with products such as driverless forklifts and AGVs, while also covering some loading, unloading and handling work. Hangcha’s X1 logistics robot extends coverage to transfers, loading, unloading, depalletizing and palletizing; Anhui Heli has also demonstrated applications of embodied and humanoid robots in logistics. The two companies address the same category of worker pool, so their headcounts should not be added together.
Why do control systems matter in welding, and how should the worker pool be understood?
Friendess Electronic’s connection to the opportunity lies in welding control systems and robotic workstations. The materials list teach-free welding, offline programming, vision positioning, weld-seam tracking and multi-machine coordination, which involve programming, positioning and operational control. Its business opportunity depends not only on demand for the robots themselves but also on adoption of these control and workstation products.
The materials put the corresponding welder pool at 10 million–30 million people in China and 40 million–80 million globally, with a focus on heavy-industry welding in steel structures, shipbuilding and construction machinery. By comparison, apparel-related roles account for 4 million–7 million people in China and 40 million–60 million globally; packaging roles account for 2 million–5 million in China and 10 million–30 million globally.
These figures describe the scope of manual jobs corresponding to the products, not equipment order volumes. In particular, forklift solutions cover only some loading, unloading and handling work. When assessing the demand opportunity, it is important to retain the distinction between job categories and specific tasks.
