Gu Rui, founding partner of Youshan Capital, shared his views on embodied intelligence during the "Physical AI and Embodied Intelligence Robot Innovation Ecosystem Exchange" event, held in Beijing on September 11, 2026, as part of the China International Fair for Trade in Services.
Gu Rui introduced his firm, stating that Youshan Capital, established in 2019, operates offices in Beijing, Shenzhen, Shanghai, and Wuhan. The investment team originates predominantly from the core group at CICC Capital, with each partner bringing 15 to 20-plus years of investment experience. The firm currently manages 50 billion RMB in assets, having made substantial allocations across technology sectors, including AI, semiconductors, aerospace, and biopharma. In the embodied intelligence field, their investments include Songyan Power and Yiai Future from the previous year, and Zhongke Fifth Discipline earlier this year, alongside companies throughout the industry chain. Their earlier investment, Huayan Robotics, successfully listed on the Hong Kong Stock Exchange recently.
When asked how he entered the field, Gu Rui explained that his journey began in 2021 with a comprehensive look at the industrial collaborative arm sector, which led to an investment in Dazu Robotics, later renamed Huayan Robotics prior to its listing. Back then, the investment logic was primarily hardware-oriented, centered on six or seven-axis force control technologies, with less clarity on the intelligent aspects.
The industry's heat started rising following the emergence of ChatGPT and around late 2023, when the release of impressive product videos from Unitree significantly boosted the sector's visibility. Gu observed a sudden leap in domestic robotics capabilities, compelling his team to track the sector closely, noting that technological iterations now occur almost every six months, driving rapid progress. Comparing the current state to autonomous driving, he parallels the evolution as moving from L1 to L2, anticipating even faster iterations ahead given the acceleration of AI development.
Acknowledging skepticism around the industry, Gu Rui recalled that while Boston Dynamics showcased remarkable demos, commercialization seemed distant, so early predictions suggested this wasn't a near-term reality. However, despite being far from AGI expectations, China's embodied intelligence has now entered its commercialization year. Multiple niche markets are starting to see practical implementation, building data flywheels that support further technological advancement, making real commercial viability attainable.
On the question of whether this is a true paradigm shift versus a recurrence of the volatility robotics has seen over the past two decades, Gu Rui is firmly convinced of the former. He pointed to the past year's breakthroughs that have transformed the technology from performance-oriented displays into genuine productivity tools. While development may still fall short of broader expectations, the trajectory is becoming increasingly clear, with continued evolution and new technologies on the horizon. He believes this is fundamentally different from the previous 20 years.
Discussing the commercial loop differences from the prior generation, Gu Rui noted the common theme of labor replacement: industrial automation previously handled certain tasks, and more intelligent robots are now needed for those that automation could not address. But the major distinction lies in what drives the new cycle. Beyond labor substitution, the new embodied robot business model centers around the data flywheel created in commercial settings and its role in enhancing generalized intelligence. This generalization, enabling skills to transfer quickly across scenarios, is the most significant value proposition of current embodied or humanoid robots. They can work in factories across different job types and extend far beyond factory settings.
Currently, the industry's first competitive step is identifying scenarios that support a closed commercial loop, but future competition hinges on broader generalization, using the same robot hardware and brain to serve more verticals. From this angle, Gu Rui believes that companies anchored primarily in industrial settings may not ultimately capture the largest market, since industrial contexts generally involve fixed environments addressing specific complex problems. Instead, he sees a bigger opportunity for embodied robots in service and new retail sectors, where diverse learning environments can foster stronger generalization capabilities, enabling these robots to cover much wider markets.
This outlook clarifies Gu Rui's enthusiasm for the retail and service sectors, where accumulated technical capabilities and data flywheels can help the robot brain learn broader generalized skills, forming a key differentiator in core competence.