Yaole Technology Develops Sensor Glove Amid Race for Tactile Data for Chinese Humanoid Robots

Nikkei reported on September 23, 2026, that Yaole Technology has developed a sensor glove amid competition over tactile data for Chinese humanoid robots.

For humanoid robots, securing training data, as well as developing the hardware itself, is seen as a factor that can determine performance. The sensor glove that Yaole Technology reportedly developed can be viewed as an effort to secure an entry point for collecting such data, making it important to watch which companies secure sources of tactile data.

The report on Yaole Technology’s sensor glove is evidence that the basis of competition in humanoid robotics is expanding beyond robot performance to include ways of collecting tactile data. However, the report is the primary source, and it does not reveal the glove’s specifications, the types or volume of data it collects, production plans or sales model. The scope of the report should therefore be assessed first. For details, please refer to the original source. The development suggests that humanoid robotics is moving beyond building machines to collecting real-world contact data and using it for training. Tactile data is harder to collect than visual data and must account for factors such as an object’s hardness, slippage and changes in contact pressure. A glove-based measurement system can be understood as a way to collect data through workers’ actions. However, publicly available information does not make clear whether Yaole Technology’s glove is intended for research or as a commercial product for mass production. Physical AI News has no related articles, but Chinese humanoid robotics companies are generally seen as seeking to develop in-house or secure control over training data, sensors and control software alongside the development of the robots themselves. In this context, companies that secure equipment for collecting tactile data early may gain an advantage in developing subsequent training models and applications. At the same time, tactile data is highly dependent on its use, and the measurement conditions required vary by action, such as gripping, pressing or turning. If data collection is not standardized, it will be difficult to apply the data across different uses. The focus of competition is likely to shift from whether sensors are present to how much data companies can collect, for which actions and at what level of precision. Points to watch include: (1) the types of signals Yaole Technology’s sensor glove captures and its accuracy; (2) how the data will be integrated into training for humanoid robots; and (3) whether the glove is intended for research or commercial use. Until these points are clear, the report only indicates that competition over tactile data is underway.