Robots once "saw" the world, are now beginning to "touch" it, and in the future, they may even "perceive" it. Driven by AI, sensor technology is leading robots into a new era of intelligent perception.
Tactile sensing, multimodal fusion, edge intelligence, and domestic substitution—behind these keywords lies a strategic emerging sector driven by both capital and technology. This issue, supported by industry data, case studies, and technological trends, will focus on five key trends in the robotics sensor industry expected to emerge in 2025.
01
Market size continues to expand: China's market annual growth rate exceeds 16%
Globally, relevant data predicts that the robot sensor market will reach US$21 billion (approximately RMB 152 billion) in 2025, of which vision, force perception and IMU (the three major types of sensors) will account for more than 70% of the market.
According to the forward-looking industry report:
In 2023, the scale of China's smart sensor market has reached 143.52 billion yuan;
It is expected to reach 157.6 billion yuan in 2024;
By 2030, it will exceed 340 billion yuan, with a compound annual growth rate of >16.5%;
Among them, the market for robot sensors is expected to exceed 40 billion yuan.

02
Amazon's "Vulcan" is released, and tactile sensors are moving from scientific research to commercial mass production
In May 2025, Amazon announced the official deployment of its latest warehouse robot "Vulcan".
<ul style="-webkit-font-smoothing: antialiased; box-sizing: border-box; -webkit-tap-highlight-color: rgba(0, 0, 0, 0); padding-left: 20px; overflow-anchor: auto; font-family: Inter, -apple-system, BlinkMacSystemFont, " segoe="" ui",="" "sf="" pro="" sc",="" display",="" icons",="" "pingfang="" "hiragino="" sans="" gb",="" "microsoft="" yahei",="" "helvetica="" neue",="" helvetica,="" arial,="" sans-serif;="" text-align:="" start;="" text-wrap:="" wrap;="" background-color:="" rgb(255,="" 255,="" 255);="" margin-top:="" 8px="" !important;"="" class=" list-paddingleft-2">Advanced core technology
The multimodal sensing system combines tactile perception, stereoscopic vision, and advanced AI algorithms to accurately sense the shape, weight, and material of goods, enabling efficient gripping of goods in irregular or confined spaces. Through deep learning model training, the gripping force can be adjusted in real time to ensure operational stability and safety.
Strong independent learning ability
The AI algorithm is based on a large amount of real-world operational data and continuously optimizes gripping strategies through reinforcement learning and accumulated experience. It has excellent adaptability in complex environments. Its autonomous learning ability improves operational flexibility and reduces operational risks caused by environmental changes. It can also intelligently identify tasks beyond its capabilities and hand them off to human operators, demonstrating the concept of efficient human-machine collaboration.
High operating efficiency
Improve working environment and safety
It is deployed in multiple warehousing centers around the world, including the warehousing nodes in Hamburg, Germany and Spokane, USA, and is mainly responsible for "high-altitude" and "low-altitude" cargo operations, reducing safety hazards in employees' heavy operations and alleviating labor shortages.
Promoting industry development
It represents the deep integration of AI innovation in smart logistics, sets a technological benchmark for the industry, consolidates Amazon's technological leadership in the industry, and provides technical reference for other companies, pushing the logistics industry into a new era of full automation and intelligence.

03
Multimodal sensing becomes popular: PolyTouch integrates vision, acoustics, and touch
A research team from MIT released a new robotic finger sensor system - PolyTouch in March this year.
The PolyTouch sensor integrates multiple tactile sensing modalities to simultaneously capture rich information about the object being touched, including its shape, material, position, and force. This multimodal design enables robots to obtain more comprehensive and detailed tactile feedback when interacting with objects, enabling them to more accurately perceive the object's characteristics and state changes.
This type of multimodal sensor provides robots with high-resolution tactile feedback across multiple timescales, enabling them to grasp fragile or irregular objects with greater stability and accuracy. It can be considered to provide robots with sensing capabilities similar to those of human fingers. Furthermore, its fine tactile feedback makes it suitable for use in applications requiring fine tactile feedback, such as medical assistance robots and rehabilitation arms.
04
Sensor Intelligence: AI + Sensing Explodes the Predictive Maintenance Market
Sensing systems not only “see” and “hear”, but also “think”.
Voliro, headquartered in Zurich, Switzerland, completed a $23 million Series A funding extension in 2025, which will be used to expand its industrial inspection drone technology. Its core product features a modular multi-sensor system integrating thermal imaging, ultrasound, and high-definition vision. Its patented tilt-rotor design enables close-range inspection of complex structures such as flare stacks and wind turbines, eliminating the need for scaffolding and reducing safety risks by nearly 80% compared to traditional methods.
Technically, drones use AI algorithms to analyze equipment vibration and wear signals, combining them with 3D modeling to generate health heat maps. Company data indicates that this solution can identify potential failures 6-12 months in advance.
At the industry level, the compound growth rate of detection solutions integrating AI and multimodal sensing is expected to exceed 25%. Voliro is promoting the transformation of industrial maintenance towards "active prevention" through the "drone + sensor + intelligent analysis" model.

05
New technologies continue to emerge: electronic skin, ultrasonic 3D sensing, and domestic substitution breakthroughs
Electronic Skin (E-Skin): A New Era of Flexible Intelligent Sensing
Electronic skin (E-Skin), a cutting-edge technology that fuses materials science and electronic engineering, is reshaping the boundaries of human-computer interaction with inspiration from bionics. Market data research shows that the global e-skin market is experiencing rapid growth, reaching RMB 78.9 billion in 2024 and projected to reach RMB 95.6 billion in 2025, a year-on-year increase of 21.1%. This growth is driven by the urgent demand for highly sensitive flexible sensing in the fields of artificial intelligence, robotics, and healthcare.
Some companies have pioneered core technologies, achieving localized mass production of independently developed flexible pressure array chips, with key performance comparable to international standards. These chips utilize nanoscale flexible substrates and self-healing materials, offering ultra-thin, high-density integration, and excellent bending resistance. This approach breaks the monopoly of foreign manufacturers, significantly reduces costs, and lays the foundation for large-scale application.
Ultrasonic 3D Sensors: Redefining the Standard for Near-Field Perception
Sonair's safety-grade ultrasonic 3D sensor, "ADAR," challenges the market dominance of lidar with its disruptive technology architecture. This sensor overcomes the accuracy bottlenecks of traditional ultrasonic arrays, achieving high-precision ranging with minimal error at close range. Furthermore, its innovative technology maintains high sensing stability in extreme environments such as rain, fog, and dust, while remaining significantly less expensive than comparable lidars.
ADAR's advantages are particularly significant in various application scenarios. In autonomous driving, it serves as the primary near-field sensing tool, filling the blind spots of lidar. In logistics and warehousing robots, the sensor can construct dynamic 3D maps in real time, guiding them to precisely avoid obstacles in complex environments. In smart homes, ADAR's contactless sensing capabilities enable human motion capture, opening up new possibilities for intelligent interaction. This technological innovation suggests that ultrasonic 3D sensors will gradually replace some of lidar's functions in future scenarios where cost and environmental adaptability are crucial.

Domestic six-dimensional force sensor: a milestone in breaking the import monopoly
In the high-end sensor sector, domestically produced six-axis force sensors are making a comeback. Companies like Qifan Sensing have successfully mastered six-axis MEMS packaging technology, significantly reducing sensor size while achieving ultra-high sampling frequencies. This technological breakthrough has brought significant cost advantages: imported six-axis force sensors were once expensive, but domestically produced products have significantly reduced their prices while achieving performance exceeding 90% of similar international products. Industry forecasts predict that by 2026, domestically produced six-axis force sensors will achieve complete self-sufficiency in core components, completely eliminating import dependence.
Domestic six-dimensional force sensors are playing a key role in Industry 4.0 and medical technology. In industrial collaborative arms, they impart "tactile intelligence" to robotic arms, significantly improving the efficiency of precision assembly. In humanoid robot joints, sensors monitor force changes across multiple degrees of freedom in real time, ensuring the stability and safety of complex movements. In remote surgical equipment, doctors receive highly accurate force feedback, breaking through spatial limitations and enabling cross-regional precision medicine. This comprehensive breakthrough, from technology to industry, marks my country's entry into the global forefront of high-end sensors.

From "sensing" to "cognition", robots are moving towards true intelligence
The development of robot sensors is moving from "isolated single-point perception" to a new stage of multimodal fusion + edge AI decision-making + data closed-loop optimization.
2025 is a "year of awakening": Vulcan awakens the robot's tactile nerves with its multimodal perception system; PolyTouch drives the robot's "multi-sensory synchronization"; Voliro promotes the implementation of AI + sensing in industrial inspections; domestic technologies gradually make up for the "bottleneck" shortcomings and move towards global competition.
In the future, robots will not only be "action machines", but also "perceptual intelligent entities", and sensors will be the "nervous system" connecting them to the world.
What we look forward to is not only the efficiency revolution brought by robots, but also a new chapter of deep integration between humans and machines and joint exploration of the unknown world.