Summary of Key Points
For AI to transition from the virtual world (where it processes voice, text, and video) to the physical world (where it performs tasks and interacts), it must overcome the challenge of "tactile perception"—and that's where electronic skin comes into play. However, the electronic skin industry faces two major hurdles: inconsistent technical approaches and difficulties in mass production. Hua Weike has overcome these issues by adopting an IDM (Integrated Device Manufacturing) model that encompasses the entire production process, making it the leader in the segment of dexterous hand electronic skin. This market segment is being redefined as a long-term investment opportunity due to the scarcity of physical data, the surge in downstream demand (as robot tactile capabilities have evolved from optional to essential), and the presence of significant data barriers. In the future, as robots begin to perform real tasks and enter households, electronic skin is poised for even greater growth.
Why Can't AI Function Without Electronic Skin?
AI has been highly effective in the virtual world—tasks like translation and content creation rely on abundant virtual data that feed large models. But for AI to perform practical tasks (such as robots picking up objects or caring for the elderly), it needs to understand the physical properties of the environment, such as the texture, temperature, and pressure of objects. This is similar to how humans use their sense of touch; without it, actions like holding a cup could result in damage or burning one's hand from hot water.
Zhu Xiaohui explains, "Observing the physical world relies on vision, but to modify it, we need tactile perception, which is the third major sensory modality after sight and hearing." Electronic skin acts as the "tactile nerves" for robots, converting physical signals like pressure and temperature into data that AI can understand. Without it, robots would act blindly and be unable to perform delicate tasks.
The Two Major Obstacles in the Electronic Skin Industry
1. Disorganized Technical Approaches
There are various types of electronic skin technologies, including piezoresistive, capacitive, and piezoelectric ones, but there is no unified standard. It's like everyone is trying to build wheels, yet each has a different design. Moreover, these technologies cannot currently achieve human-like tactile sensitivity (simultaneously detecting pressure, temperature, and humidity), leaving it uncertain which approach will prevail.
2. From Laboratory to Mass Production
The transition from laboratory development to mass production is problematic. Samples can be created in the lab, but scaling up is difficult: either the right materials are unavailable, the equipment is insufficient, or the yield rate is low (with outsourcing only reaching 60%-80%). Zhu Xiaohui illustrates, "The device and algorithm contribute 90% of the success, but materials and equipment account for just 20%, which ultimately prevents mass production."
How Hua Weike Became the Leader in This Segment
Hua Weike's solution is its IDM model, where it handles every aspect of production—from materials to equipment, components, to algorithms. This approach is unprecedented in the industry and offers several significant advantages:
- Fast Iteration: Issues can be addressed immediately without waiting on suppliers, resulting in faster response times.
- High Yield Rates: By controlling all stages, Hua Weike achieves yield rates of over 99% for most products.
- Lower Costs: Long-term cost savings are possible due to complete self-sufficiency in the production chain.
- Flexibility in Technology Transition: If a technology becomes obsolete, Hua Weike can quickly switch to another, avoiding bottlenecks.
As a result, Hua Weike supplies electronic skin to leading manufacturers like Yinzhou Robot and Zhiyuan. By 2025, it expects to produce over 10,000 units of dexterous hand electronic skin, accounting for more than 70% of the market share, thus establishing itself as the industry leader.
The Investment Value of This Segment
The potential of this segment goes beyond just hardware; it also includes data and algorithms:
- Surging Downstream Demand: The penetration rate of robot tactile capabilities has increased from 20%-30% last year to over 60% this year, with further growth expected. Hua Weike's order volume has grown from 10,000 units to 50,000 units this year, reflecting this trend.
- High Data Barriers: Tactile data lacks a unified standard like visual (images) or auditory (voice) data. Different sensors interpret the same physical input (e.g., 0.5 Newtons of pressure) differently, making it essential for algorithm companies to collaborate with hardware manufacturers to obtain meaningful data. This allows hardware providers like Hua Weike to expand into data and analytics services (e.g., providing solutions for sleep technology companies).
- Re-evaluation by Investors: In the past two months, at least nine tactile technology companies have raised funding. Investors are no longer viewing electronic skin manufacturers merely as hardware producers but considering the additional value that data and algorithms bring, opening up significant valuation opportunities.
Future Growth Drivers
Two major events will drive further growth in the electronic skin industry:
1. Robots Taking on Real Work in Factories: As robots begin to work in factories, their tactile capabilities will become more critical.
2. Robots Entering Homes: Robots used for tasks like caring for the elderly and household chores will require advanced tactile systems. This will lead to a shift from limited applications (e.g., fingertips) to comprehensive coverage of the robot's entire body, potentially evolving into "bionic skin."
Conclusion
Electronic skin represents the final step for AI to fully integrate into the physical world. Although the industry is still in its early stages, Hua Weike's IDM model has demonstrated the potential for breakthroughs. With the surge in downstream demand and the increasing value of data, this segment is likely to attract long-term investment interest. After all, enabling AI to truly interact with the physical world is no small feat.