This article only summarizes the latest fundamentals of companies and the industry, and does not recommend buying or selling companies at this time.This article does not provide guidance for individual stock operations; investing involves risks, and one should proceed with caution.
Since the beginning of this year, the rotation pattern of the robotics sector
The PE (TTM) of the automotive parts industry is expected to drop from over 50 to around 20 times from 2022 to 2024. Against the backdrop of the large robotics industry, the valuation center has entered a long-term upward channel.
Structurally: Large companies – performance + core progress in the industry; medium-sized companies – performance + their own progress in the robotics industry chain; small companies – their own progress in the robotics industry chain.
From October 2024 to February 2025, large companies like Sanhua, Topband, and Hengli will generally see significant increases when there are new developments in the North American or domestic industries;
From December 2024 to April 2025, medium-sized companies like Yinlun and Zhongding will experience major rises under stable performance and new developments in the robotics industry;
From March 2025 to June 2025, small and medium-sized companies like Rongtai and Hengbo will welcome major upward trends with new developments in the robotics industry.
Leading companies have adjusted sufficiently, and valuations are at the bottom; we are generally bullish on the robotics sector.
In terms of valuation, the overall component valuation is around 28 times, which is only a 20%-30% increase compared to the bottom at the end of 2024, with clear industry trends and significant room for valuation improvement.
The eve of a humanoid robot explosion
On August 21, the first partner conference of Zhiyuan Robotics was held in Shanghai, featuring four sub-forums, eight commercial scenario exhibition areas, and the AgiBot Night technology party, showcasing Zhiyuan Robotics’ comprehensive layout in “products, technology, business, ecology, capital, and team”.
Launch of Lingxi X2-W: A wheeled dual-arm robot prototype designed for “operational intelligence”, expected to become a benchmark for the next generation of embodied intelligent operational robots. It features an omnidirectional mobile chassis, high-degree-of-freedom dual arms, bionic wrists, compact storage (occupying less than 0.5㎡), dual power system switching, dexterous three-finger hands with tactile feedback, omnidirectional perception systems, abundant computing power edge computing units, and low cost.
“1+3” full-stack technology strategy: Based on the robot body, it builds three core capabilities: motion intelligence, interaction intelligence, and operational intelligence, creating a self-evolving general embodied intelligent base, with three product series – Expedition, Sprite, and Lingxi.
Technical open-source: Robot middleware AimRT, a million real machine dataset, and the launch of the first embodied intelligent operating system “Lingqu OS”.
Commercial ecosystem construction: Implementing an “integration strategy” covering eight scenarios including explanation reception, entertainment performances, industrial intelligence manufacturing, logistics sorting, security inspections, commercial cleaning, data collection training, and scientific research education, while building a layered distribution system.
Launch of the “Zhiyuan A Plan”: The first entrepreneurial acceleration plan focused on the embodied intelligence industry chain, aiming to incubate 50+ high-potential early-stage projects and create a trillion-level industrial ecosystem within three years.
The Zhiyuan full-stack system architecture is clearer
One body, three intelligences: Robot body + core capabilities of motion intelligence, interaction intelligence, and operational intelligence.
Motion intelligence focuses on sim2real reinforcement learning to achieve adaptive walking in complex terrains; interaction intelligence must understand and engage in conversation, with a multimodal dialogue response speed of one second; operational intelligence achieves a closed loop from grasping to fine operations through real machine reinforcement learning.
The positioning of product series is clearer: The Expedition series full-size humanoid robots focus on explanation reception and entertainment performances; the Sprite series integrates native data collection and pushing, paired with platforms like Genie Studio, suitable for various industrial and commercial scenarios; the Lingxi series, at 1.3m in size, covers scenarios such as entertainment performances, store reception, and scientific research education. The X2-W, a wheeled dual-arm robot, targets operational intelligence.
Scenario implementation: Focused on eight scenarios including explanation reception, entertainment performances, industrial intelligence manufacturing, logistics sorting, security inspections, commercial cleaning, data collection training, and scientific research education, solutions have been launched and scaled applications across multiple industries.
More platform support for the industrial ecosystem: 1) The Lingchuang platform lowers the threshold for secondary development, making it easier to customize robot behaviors; 2) The Zhiyuan A Plan aims to incubate 50+ high-potential early-stage projects, with the first entrepreneurial camp officially launched on August 21, 2025, targeting global robotics startups and developer teams; 3) Launch of partner policies to provide comprehensive support, co-building a “technology – product – scenario” collaborative ecosystem and sharing industrial dividends.
This supply chain conference also featured new faces, such as Leadshine, Digital China, Laimu, and Weishijiajie, with the Zhiyuan chain ecosystem continuing to solidify. Leadshine was invited to attend the Zhiyuan Robotics partner conference.
On the evening of August 19, Yushu Robotics released a new product preview poster, showing a robot height of 180cm, with 31 degrees of freedom (96*2+3+7*2+2), and the words Agile and Elegant below.
Humanoid robot stock selection strategy: Look for momentum under three necessary conditions (management’s determination, past performance verification, future scenarios), buying small, new, low, leading, and application scenarios!