Paxini Starts A-Share IPO Tutoring, Valued Above 10 Billion Yuan, Uses Sensor Gloves for Data

Another robotics company has begun the process of listing on China’s A-share market. According to the website of the China Securities Regulatory Commission, Paxini AI Technology Co., Ltd. has submitted a filing report for IPO tutoring and listing to the Beijing Regulatory Bureau of the CSRC. The filing is the latest sign that Chinese robotics companies are accelerating their push into public markets.

The report says Paxini signed a tutoring agreement with Guotai Haitong Securities on September 15. The company is registered in Haidian District, Beijing, with registered capital of 14.233 million yuan.

Paxini’s legal representative and controlling shareholder is Xu Jincheng, who controls 37.07% of the company’s shares. This indicates relatively concentrated ownership and continued strong control by the founding team over the listing process.

Paxini’s approach to data collection differs from that of many peers. The industry commonly uses teleoperated robots to grasp objects and collect data, while Paxini has people wear sensor-filled gloves to work directly. The difference concerns both the source and quality of data.

Teleoperated robots generate action data from robots, whereas sensor gloves capture real data from human operations. When a human hand grasps an object, its force, angle and micro-adjustments reflect fine control formed through long evolution. For robots to learn these actions, collecting data directly from human hands may be more efficient than repeated trial and error.

Paxini itself develops tactile sensors. The sensors packed into its gloves record tactile information during human operation. This data can include contact force, pressure distribution and timing changes, and it can be used to train robots’ dexterous manipulation capabilities.

Data quality is critical for embodied AI training. Robot learning often depends on large amounts of demonstration data, and the closer that data is to real human manipulation, the more likely it is to train policies with stronger generalization. Human hands can perform delicate tasks that are difficult to script for a robot. By capturing human demonstrations, Paxini can build datasets that reflect real-world variation in object shape, weight and surface conditions.

Tactile sensing is Paxini’s core focus. Founded in Shenzhen in 2021, the company concentrates on various types of tactile sensors. Founder and CEO Xu Jincheng graduated from Waseda University in Japan, and CTO Zhang Hengdi comes from the French National Centre for Scientific Research. The company has raised more than 4 billion yuan cumulatively, and its valuation has exceeded 10 billion yuan.

In the embodied AI supply chain, tactile sensing is becoming an independent segment. For robots to perform fine manipulation, they need both vision and touch: vision tells a robot where an object is and what it looks like, while touch tells it whether the grasping force is sufficient, whether the object is slipping, and what material its surface is made of.

Tactile information can also help robots avoid crushing fragile objects and handle deformable or slippery items. In practical terms, touch complements vision in tasks such as assembly, packaging and household service, where contact-rich interaction is common.

Paxini enters from tactile sensors, uses sensor gloves to collect data, and then applies tactile capabilities to products such as dexterous hands, humanoid robots and electronic skin. Compared with pure whole-machine companies, this business structure, based on sensors and core components, has already achieved batch shipments and has clearer revenue sources. The company is positioning itself in the embodied AI supply chain as a sensor provider that also supplies data. That dual role could help it serve robot makers while building its own data advantage.

The CSRC’s recent IPO review of humanoid robot companies focuses on recurring revenue, loss improvement and genuine innovation capability. Paxini’s tactile sensor products have already entered the batch shipment stage, and this business structure may make it easier to meet regulatory requirements than pure whole-machine companies. For investors, the key question is whether tactile sensing can scale from niche components to a standard part of humanoid robots and dexterous hands. Paxini’s early shipments suggest it is testing that path.