Embodied-AI robot tactile technology developer Yuansheng Technology has raised more than 100 million yuan in a new funding round. The investors include financial backers 5Y Capital, Linear Capital, Houxue Capital and Xichen Capital, alongside strategic investors RoboSense, TOP Group and Shangfu Electronic Technology, respectively a listed lidar company, an automotive Tier 1 supplier valued at more than 100 billion yuan, and an automotive intelligent-driving sensor company preparing to go public. Zhao Yuekai, CFO of European embodied-intelligence unicorn Agile Robots, also invested in a personal capacity.

The company previously secured angel and Pre-A funding from Professor Li Zexiang, Keli Sensing, NOVOSENSE, Qigao Capital and Hengxin Huaye.

Putting the two groups of names side by side reveals more than the headline figure of “over 100 million yuan.”

The Money Is Coming From the Automotive Supply Chain

Founded in January 2025 by an Oxford University PhD, Yuansheng Xianda Technology (Shenzhen) develops tactile infrastructure for embodied intelligence. It has built a full stack spanning tactile sensors, tactile data, tactile intelligence and scaled manufacturing.

What this stack needs to reach the market is clear. RoboSense brings mass-production experience in automotive-grade optical sensing; TOP Group contributes the system integration and assembly capabilities of an automotive Tier 1 supplier; Shangfu Electronic Technology provides automotive-grade manufacturing expertise in intelligent-driving sensors; Keli Sensing and NOVOSENSE specialize in force sensing and automotive-grade chips, respectively; and Li Zexiang represents the incubation ecosystem for robotics and hard technology.

Bringing these shareholders together with a tactile technology company founded only in 2025 points to one conclusion: the contest is no longer about whether a prototype can be built, but whether automotive-industry standards can deliver consistent products at scale.

Three Products Cover Hands, Bodies and Data

The company has organized its product portfolio into three layers, SENSE, CAPTURE and LEARN, each aimed at a different type of buyer.

The SENSE layer comprises tactile components. Its proprietary TACTA multidimensional tactile sensor combines high force-field resolution, native multimodality and multidimensional force sensing. Development is focused on lowering integration barriers for customers, particularly by solving the engineering challenges of miniaturization and high sensing density. The goal is direct compatibility with dexterous hands, grippers and robotic arms in a range of form factors. For large-area collision detection and interaction across a robot’s torso, the company has also developed HEXSKIN, an electronic skin that conforms to curved surfaces and extends sensing from the end of the hand to the entire body.

This layer sells hardware to robot manufacturers and dexterous-hand companies.

The CAPTURE and LEARN layers move one step further downstream. To address the shortage of real-world robotic data faced by large-model companies, Yuansheng launched Magic Glove, a data-capture glove. As data collectors wear the glove to grasp and manipulate objects naturally, it synchronously records hand trajectories and contact forces, converting them directly into real-world datasets for training large models.

This layer sells an entry point for data collection to algorithm teams that need real-world robotic data to train their models. It also sets Yuansheng apart from most tactile technology companies: the business is not entirely dependent on robot manufacturers ramping up production.

The Automotive-Grade Production Line Is the Real Asset Behind This Round

The sentence most likely to be treated as background in the funding announcement is, in fact, its most substantive claim.

Yuansheng Technology and Shangfu Electronic Technology have jointly built a dedicated production line for multidimensional piezoresistive tactile sensors, transplanting the automotive industry’s most stringent quality-management systems, fully automated manufacturing processes and end-to-end traceability mechanisms into tactile sensor production. In August this year, the company officially announced the start of production at what it calls the world’s first automotive-grade mass-production line for multidimensional tactile sensors.

The significance of the production line lies not in the claim that it is the “world’s first,” but in how it changes the way the company is assessed. At the prototype stage, engineers can manually calibrate each unit and fine-tune parameters to meet performance targets. Once orders scale, however, automated testing, parameter compensation and quality traceability must be embedded in the production process, or consistency across a single batch cannot be maintained.

With the automotive-grade production line in place, the company says it has moved beyond small-batch sample validation, achieved reliable automotive-grade mass production, and begun shipping designated-volume orders to several prominent dexterous-hand, humanoid-robot and data-capture customers in China and overseas.

“Designated supplier” and “volume shipments” describe two different milestones. Designation means the product has entered a customer’s bill of materials; volume shipments mean production capacity and quality control have been tested under real operating conditions.

How the Funding Will Be Used

The new funding will support the development of next-generation tactile products and foundational models, expansion of the automotive-grade production line, and international market growth.

The three priorities correspond to three stages of growth: next-generation products will determine whether Yuansheng can keep pace with changing dexterous-hand designs; foundational models will determine whether its data business can stand on its own; and production-line expansion will determine whether it can fulfill orders as they arrive. International expansion comes last, suggesting the first two steps remain unfinished.

The company says it aims to be the first to cross the mass-production threshold using automotive-industry standards and deliver highly consistent products to customers at scale. Put simply, the decisive question in tactile sensors has shifted from “Can it be made?” to “Can it be made the same way every time?”