Shenzhen-based firm secures more funding amid growing demand for automotive-grade tactile sensing technology in embodied AI.
By Zan Zhu
Chinese tactile sensing startup Yotlive has raised a new Pre-A+ funding round led by Dinghe Gaoda, as investors increasingly bet that physical interaction data will become a key component of the next generation of artificial intelligence models.
The round also attracted participation from listed automotive supplier Changshu Automotive Trim (603035.SH) and game developer Archosaur Games (Zulong Entertainment) (9990.HK). Existing investor Yundao Capital acted as the company’s long-term exclusive financial adviser. Financial terms were not disclosed.
The latest financing follows an angel investment from tech giant Xiaomi (1810.HK) and a Pre-A round backed by Ecovacs Robotics-affiliated Yinfeng Fund, Silicon Harbor Capital, and Chuhui Capital.
The fundraising comes as investor enthusiasm for embodied AI has cooled from last year’s peak, while attention has shifted toward a more fundamental challenge: the shortage of real-world physical interaction data needed to train AI systems that operate outside digital environments.
Industry participants increasingly believe that while vision-language-action (VLA) models – which map what the robot sees and hears directly to a physical action – have advanced rapidly, the development of AI world models — systems designed to understand and predict how the physical world behaves — is constrained by a lack of large-scale datasets describing how humans interact with objects through touch, pressure and movement.
From automotive sensors to embodied AI
Founded by Lü Liyun a former executive at international automotive electronics giant Harman, Yotlive was originally built around automotive-grade textile-based tactile sensors rather than conventional rigid sensor technology.
Lü previously led multimodal sensor platform development at Harman, serving automakers including Maserati and Porsche. Co-founder Zhou Xiao specializes in integrated circuits and intelligent sensing, while another co-founder, Luo Zheng, developed multimodal sensor fusion algorithms that have been deployed in hundreds of thousands of vehicles.
The company began with an attempt to develop a smart yoga mat in 2015 but found existing sensor technologies either too fragile or too rigid for continuous use.
That led the founders to focus on textiles as sensing surfaces. Instead of attaching sensors onto objects, Yotlive weaves sensing capability directly into fabrics such as gloves, vehicle seats, mattresses and insoles, allowing data to be collected without altering normal human movement. The approach aims to capture subtle physical interactions—including grip force, posture, balance and tactile feedback—that conventional visual systems cannot measure.
The company’s first commercial customer was Chinese AI company iFlytek (002230.SZ), which used the technology in pressure-relief mattresses designed for elderly care.
Its biggest commercial breakthrough came in the automotive sector. Beginning in 2018, Yotlive developed door panel and seat pressure sensing systems capable of meeting stringent automotive qualification standards covering durability, reliability, and environmental performance.
The company passed its first automotive certification in 2019 and qualified seat sensing systems in 2022. By 2025 it had become a core supplier of intelligent seat pressure systems for a leading luxury vehicle manufacturer while securing production contracts with several major automakers.
Chief executive Lü said the automotive business deliberately served as the company’s highest technical benchmark.
“Entering the automotive industry allowed us to refine the technology under the toughest standards before expanding into other applications,” she said.
Data rather than hardware
Yotlive expects orders to increase tenfold in 2026, with almost half of the incremental growth coming from embodied AI customers. Many robotics developers have approached it because increasingly sophisticated robotic hands still lack the physical interaction data needed for precise manipulation, it says.
Vision systems can identify object positions but cannot accurately measure pressure, slippage or contact quality. Simulation can supplement training but cannot fully replace real-world physical interaction.
This year Yotlive plans to launch its first commercial data glove, a textile glove embedded with pressure sensors that resembles an ordinary glove while recording detailed tactile information.
The company ultimately hopes such products become low-cost, widely deployed data collection terminals that continuously generate standardized tactile datasets for AI model training.
Yotlive has already established partnerships with leading robotics companies and consumer wearable manufacturers. It has also built a roughly 7,500-square-meter factory in Jiaxing that covers the production process from specialized yarn through completed sensing networks.
The company argues that manufacturing expertise, rather than sensor hardware alone, represents its competitive advantage.
Over the longer term, management sees Yotlive evolving into a tactile data platform rather than simply a sensor supplier.
Hardware sales provide cash flow and validate applications, but executives believe the larger opportunity lies in building standardized datasets that AI developers can access through calibrated, cleaned and labeled interfaces.
According to Lü, customers are increasingly asking not whether tactile hardware exists but whether the resulting data is ready for AI training.
She said the long-term competitive advantage would come from creating a self-reinforcing data network in which broader hardware deployment generates more data, improving AI models and attracting additional customers.
Strategic investors back long-term vision
Yotlive’s financing history mirrors the changing priorities of China’s AI investment landscape.
Xiaomi invested in the company at the angel stage in late 2022, when Yotlive had completed automotive qualification but had yet to secure large-scale production orders. In late 2025, Yinfeng Fund led a Pre-A round worth nearly 100 million yuan ($14.8 million), with proceeds earmarked for technology development, overseas expansion and mass production.
The latest round brings in Dinghe Gaoda, whose portfolio includes autonomous driving and sensing companies such as Hesai (2525.HK) (HSAI.US) and Momenta (6800.HK), alongside strategic investors with automotive and overseas industry resources.
Changshu Automotive Trim said growing overseas demand for occupant classification systems creates opportunities to export China’s automotive-grade sensing technology to international vehicle manufacturers.
For investors, the sector remains immature, but evaluation criteria are becoming clearer: proven commercial deployment, scalable manufacturing, competitive costs and the ability to generate standardized physical interaction data that can support future Physical AI models.
Source:
ChinaVenture
