Wang Xingxing, founder and CEO and CTO of BotSchool, stated during his speech at the main forum of the 2026 World Robot Conference (WRC2026) that embodied intelligence is approaching an industrial tipping point similar to the "ChatGPT moment." It is expected to take as few as 2 to 3 years, or up to 5 to 10 years in a slower scenario, to achieve breakthroughs.
Just one day before the speech, BotSchool had just listed on the Science and Technology Innovation Board of the Shanghai Stock Exchange. On its first trading day, the closing price rose by 460% compared to the issue price. Wang Xingxing stated that the standard for determining when embodied intelligence enters an industrial boom is not a single scenario demonstration, but rather whether robots can complete about 80% of tasks through voice or text commands in 80% of unfamiliar environments.
Wang Xingxing pointed out that the biggest technical bottleneck of humanoid robots currently lies in the alignment between AI models and the real physical world. Robots can achieve high success rates in fully trained fixed scenarios, but when objects or environments change slightly, task completion rates drop significantly. Especially in the "last few centimeters" or even "last few millimeters" of operation, issues such as tactile feedback and action errors still limit the generalization ability of robots.
To address this challenge, BotSchool demonstrated its pre-researched "Physical AI Robot Self-Evolution V1.0" technology roadmap. This system plans to use AI large models to automatically retrieve research results, generate control code, and form a continuous iteration loop through simulation training, real machine testing, and AI and human evaluation feedback. BotSchool believes that the capabilities of foundational models, accumulation of multi-source data, scale of real robot deployment, and skill accumulation will become key factors driving robot self-evolution.
In his speech, Wang Xingxing also reviewed BotSchool's development journey from quadruped robots to humanoid robots. The company's humanoid robot G1 has become an industry representative product and participated in projects like the Spring Festival Gala's "WuBOT." This year, BotSchool also launched products such as the manned exoskeleton and the wheeled-legged robot As2-W, and continues to explore applications of robots in factory, home, and outdoor scenarios.
Wang Xingxing said that to achieve true large-scale deployment, robots still need to solve issues related to efficiency and generalization. Currently, robots can perform some simple tasks, but for new tasks, retraining is often required. With the improvement of AI model capabilities and the continuous accumulation of real-world data, embodied intelligence is expected to accelerate into the industrialization stage.
As humanoid robots move from experimental demonstrations to commercial applications, data, models, hardware, and self-evolution capabilities are becoming core elements of industry competition. BotSchool's proposed technology roadmap reflects the industry's shift from "manufacturing robots" to "enabling robots to continuously learn and evolve."