In a textile recycling workshop in Zhangjiagang, Jiangsu Province, a production line that once required 30 workers toiling for eight hours to process 15 tons of used clothing now operates with only two staff members. Behind this dramatic transformation is an automated, AI-driven textile sorting line quietly working at incredible speeds. By integrating artificial intelligence and hyperspectral imaging, this Chinese innovation addresses one of the global fashion industry’s most pressing environmental bottlenecks, paving the technical foundation for a true circular economy.
Developed by Hangzhou-based Databeyond Technology, this advanced sorting system identifies the material and color of waste textiles with unprecedented precision. The equipment can accurately detect blend ratios and automatically separate components such as cotton, polyester, and nylon. Hyperspectral imaging, the core sensing modality used, captures spatial and spectral information simultaneously, enabling the non-invasive and label-free characterization of chemical and biological properties.
The global significance of this technology was recognized last October when Time magazine named it one of the Best Inventions of 2025. Databeyond Technology stands as the only Chinese high-tech enterprise to receive this honor in the recycling category. By simultaneously improving sorting accuracy and efficiency, the technology has catalyzed a massive upgrade across the entire industrial chain.
The Bottlenecks of Traditional Textile Recycling
Mo Zhuoya, the founder of Databeyond Technology, an AI-driven innovator in the solid waste sorting sector, shifted her focus to textiles after witnessing a harsh reality during her research. At a waste recycling station in Guangdong Province, she observed mountains of discarded clothing emitting a pungent odor, while workers hunched over, manually sifting through piles mixed with cotton, linen, wool, buttons, and zippers. Despite the grueling labor, both efficiency and accuracy remained lackluster.
“This is not an isolated case; it is a ‘common disease’ in the textile recycling industry,” Mo explained to the China Environment News. Having spent years in the sector, she realized that sorting is the critical, often overlooked link determining the industry’s overall efficiency.
The quantitative data highlights the scale of the problem:
-
In 2024, China generated approximately 36.4 million tons of textile waste, representing about 26 percent of the global total.
-
However, only 5.15 million tons of this massive volume were successfully recycled.
-
Currently, over 80 percent of sorting in the traditional textile recycling industry relies entirely on manual labor.
-
A skilled worker can only sort about 100 kg of used clothing per hour.
-
Labor costs account for over 40 percent of total operating expenses.
Furthermore, clothing composition has become increasingly complex. Blended fabrics and various chemical dyes make textiles harder to identify than almost any other industrial raw material, causing traditional manual methods to fail in meeting the high-purity demands of downstream recyclers.
The “Eye, Brain, and Hand” Collaboration
To solve these persistent bottlenecks, Databeyond developed an AI-driven hyperspectral optical sorter that functions with collaborative “eyes, brains, and hands,” as described by Liu Jiahua, Market Manager at Databeyond Technology.
-
The Eyes: The system utilizes a multi-sensor vision system combining hyperspectral and laser sensors. As materials move along a conveyor belt at a rapid 4 meters per second, the system captures the shape, color, and material characteristics in real-time. By creating a “spectral fingerprint,” it can identify complex blends, such as a garment made of 90 percent polyester, 8 percent cotton, and 2 percent spandex.
-
The Brain: Because spectrum identification alone cannot support complex industrial scenarios, the company built a massive spectral database of solid waste materials. This database is integrated with AI algorithms to perform instant, real-time analysis ensuring accuracy, stability, and scalability.
-
The Hands: Handling lightweight materials like shredded fabric or garment trimmings at high speeds is physically challenging due to airflow. To combat this, the team introduced an innovative aerodynamic design featuring a positive pressure system above the conveyor belt. This acts as an invisible hand, keeping light materials flat and stable. Once identified, high-pressure air nozzles eject the target material from the belt in milliseconds with over 98 percent accuracy.
Environmental Benefits and Global Expansion
The environmental impact of this intelligent sorting technology is profound. Lin Xianping, Executive Deputy Secretary General of the China Urban Expert Think Tank Committee, noted that recycling significantly reduces the extraction pressure on raw materials like cotton and petroleum. It also mitigates carbon emissions, water pollution, and the massive land use associated with landfills.
According to official data from the UN Environment Programme, for every single ton of old clothing recycled:
-
3.6 tons of carbon dioxide emissions are prevented.
-
0.5 tons of crude oil are saved.
-
20 tons of water are preserved.
The global demand for Databeyond’s intelligent equipment is growing robustly. The company’s overseas market has already expanded beyond Japan, the Republic of Korea, and Southeast Asia, now reaching India, Mexico, Brazil, Uganda, and countries in Central Asia. Looking ahead, the company plans to deepen its AI integration, expanding into waste plastics, film, and construction waste to accelerate the intelligent era for the global recycling industry.
Official Source and Reference: