A worker works at a textile sorting facility featuring Databeyond’s Fastsort-Textile AI sorting machine in Zhangjiagang, eastern China’s Jiangsu province, on March 20, 2026.
Ng Han Guan/AP
ZHANGJIGANG, China (AP) – In an industrial park in Zhangjiagang, a small city on China’s east coast, large, humming machines collect and sort piles of used clothes.
What’s the novelty? Use artificial intelligence to quickly sort by composition and get a glimpse of how AI can play a role in reducing the impact of synthetic textile waste.
Article continues below this ad
The Fastsort-Textile machine, named one of Time magazine’s best inventions of 2025, was created by DataBeyond, a Chinese AI recycling company founded in 2018.
Mo Zhuoya, CEO of DataBeyond, said, “We will be able to make the most of textile waste and reduce the amount of incineration, which will greatly contribute to resource recycling.”
Synthetic fibers are derived from fossil fuels and are a popular low-cost choice in fashion production. Together, they account for about 70% of the world’s textile production, according to a report by Circle Economy, an Amsterdam-based nonprofit that analyzes ways to reduce textile waste.
Textile waste is a major global pollutant, with China being the main source. China leads global textile exports with $142 billion, more than double that of the European Union, according to the World Trade Organization’s 2025 Key Insights and Trends report.
Article continues below this ad
Fastsort-Textile is only used in one location in China. That is Shanhesheng Environmental Technology Ltd., a textile recycling facility in Zhangjiagang city that installed the machine in 2025.
This device uses an AI scanner to read the composition of fibers and can separate and recycle each fiber.
Fastsort-Textile sorts 100 kilograms (220 pounds) of clothing in two to three minutes. In comparison, it takes one worker approximately 4 hours to perform the same task. According to Shanhesheng’s analysis, the machine can process two tons per hour, but it takes two people two days and is less accurate.
Measuring 5 by 2 meters (16 by 6.5 feet), the AI scanner operates on a series of conveyor belts. A worker places a bundle of fibers on a belt and passes it through a scanner, which emits a sharp hissing sound as it reads the composition of the fibers. The live video feed shows the scanner-side readings.
Article continues below this ad
It takes less than a second to accurately read the material composition of one item set according to the customer’s desired benchmarks.
After the scanning process, the fibers are transported to the nylon and polyester sorting area for recycling. Substandard items are sorted into separate areas, primarily for incineration or landfilling, where textile contamination is the most damaging.
“This kind of thing saves labor costs and saves time. When people sort materials, they can’t accurately distinguish whether it’s 80% or 90% polyester. This machine rarely makes mistakes,” said Cui Peng, sales manager at Shanhesheng.
Article continues below this ad
Previously, up to 50% of processed textile products were considered non-recyclable and were sent to landfill or incinerated. Sales director Li Bin says the Fastsort-Textile machine reduces that number to 30%.
“Currently, machines are already capable of sorting, but humans have limited energy,” he says. “Humans cannot work 24 hours a day, so robots may eventually take over that role. The ultimate goal is a ‘dark factory’ where robots operate 24 hours a day.”
