The recent demonstration of the H2-X humanoid robot by Unitree Robotics ended in a malfunction when the robot stumbled during a side flip, resulting in the detachment of part of its leg. This incident occurred on 22 October at the 2026 World Physical AI Conference and Unitree Partner Ecosystem Summit, which took place in Hangzhou, located in East China’s Zhejiang Province. The event showcased advancements in robotics and artificial intelligence, but the mishap raised questions about the reliability of such technologies.

The H2-X robot was introduced to the public just a day prior to the incident, on 21 October. Footage released by Unitree demonstrated the robot’s capabilities, including actions such as kicking a car door, lifting a vehicle, and executing a flip from a height of approximately four metres. The robot is reportedly intended for security and emergency rescue operations, highlighting its potential applications in critical situations.

Following the incident, the reactions on Chinese social media platform Sina Weibo were varied. While some users expressed disappointment over the failure, others acknowledged that such errors are part of the technological development process. The mixed responses indicate a broader conversation about the expectations and realities of robotic performance.

Social Media Reactions and Discussions

The failed demonstration sparked a range of discussions among users on social media. Some commentators pointed out that while the robot is not yet mass-produced, minimising errors is crucial for future iterations. Others argued that glitches are an expected part of the evolution of technology, suggesting that the focus should be on improving the underlying algorithms and mechanical designs.

One user highlighted the need for enhancements in the robot’s structural integrity, while another emphasised the importance of considering material strength and backup systems in engineering design. This sentiment was echoed by several users who called for comprehensive testing to ensure that the robots meet safety and quality standards before being deployed in real-world scenarios.

Unitree Robotics’ founder and CEO, Wang Xingxing, addressed these challenges during a conference held on the same day as the incident. He discussed the potential for robots to achieve a significant leap in their capabilities, which he likened to a “ChatGPT moment” in the near future. Wang expressed optimism about the advancements in physical artificial intelligence, predicting breakthroughs within the next one to three years.

Previous Incidents Involving Humanoid Robots

This incident is not isolated, as there have been previous occurrences of humanoid robots malfunctioning during public demonstrations. In May, a humanoid robot in Shenzhen experienced a failure while performing a dance routine to Michael Jackson’s “Billie Jean.” During the performance, it stumbled over a step and briefly regained balance before falling again, prompting a human assistant to remove it from the stage.

Another notable incident took place during a martial arts demonstration in Xinjiang Uygur Autonomous Region, where a robot fell as children approached to join the performance. The footage showed the robot attempting kung fu kicks, leading nearby performers to step back cautiously. Such incidents highlight the unpredictable nature of robotic behaviour in dynamic environments.

In a separate event in Shaanxi Province, a Unitree robot accidentally struck a young boy in the face while swinging its arms during a demonstration. These occurrences raise important questions about safety and the need for rigorous testing of robotic systems before public demonstrations.

Long or Short, get news the way you like. No ads. No redirections. Download Newspin and Stay Alert, The CSR Journal Mobile app, for fast, crisp, clean updates!

App Store – https://apps.apple.com/in/app/newspin/id6746449540

Google Play Store – https://play.google.com/store/apps/details?id=com.inventifweb.newspin&pcampaignid=web_share