The secondhand garment market has been growing faster than fast fashion for years, yet its sorting processes remain stuck in manual labor. A machine that can identify fabric, style, and wear level in seconds is no longer a concept—Swedish machinery maker Eton Systems has launched its production-ready AI sorting module, aUPS, at Texprocess in Frankfurt. The module rides on Eton’s well-established UPS overhead transport system, adding AI-powered vision units at key nodes along the rail network. This transforms the system from a mere conveyor into a combined sorter and quality inspector. The system can recognize fabric composition, color, style tags, and the condition of accessories like buttons and zippers. Once a used T-shirt is hung on the carrier, its fate is instantly determined: resale, cleaning and refurbishment, or fiber recycling.
The need for automation is driven by sheer volume. Global secondhand apparel transactions are expected to surpass $100 billion within five years, while high return rates in e-commerce and fast fashion add further pressure. Manual sorting tops out at around 200 items per hour per worker, with consistency issues. Labor costs in developed markets keep rising, and margins in the resale trade are thin. Without automation, the economics of the entire recycling chain break down. aUPS promises to lift throughput from hundreds to thousands of items per hour, reducing reliance on skilled labor.
For sorting centers, the capital outlay is significant but the long-term savings in labor and improved accuracy can shorten payback periods, especially in labor-short Europe. For fast-fashion brands, efficient sorting directly supports ESG targets by shortening the time used garments spend in inventory before being resold or recycled. For fiber recyclers, sorting precision determines the purity of recovered materials—AI can better separate pure cotton from polyester blends, boosting the quality of recycled fibers.
Still, the aUPS is fresh from a tradeshow debut with limited real-world deployment. Procurement teams must evaluate total cost, integration complexity, and the AI model’s ability to generalize across different brands and levels of wear. Error rates in actual conditions remain to be proven with batch data.
Practical recommendations
#### For buyers (sorting centers/recyclers)
- Prioritize bottleneck analysis. If manual sorting accounts for over 30% of operating costs, automation ROI may be shorter than expected.
- Request test data from multiple batches and sources, focusing on fabric recognition accuracy and mis-sort rates.
- Choose modular systems that can integrate with existing transport lines to minimize downtime during upgrade.
#### For foreign trade companies (used garment exporters)
- Monitor tightening EU regulations on textile waste exports. Poor sorting accuracy may lead to higher compliance costs.
- Coordinate technical specifications with European sorting centers that adopt AI systems—adjust packaging and labeling for automated handling.
- Incorporate sorting data into pricing models. More precise classification commands higher resale prices, allowing exporters to optimize procurement pricing.
The automation of used garment sorting is not a question of if, but when and how. Eton’s aUPS is the first domino; more machinery makers will enter this niche. For the entire textile circular economy, sorting efficiency is the key to closing the loop from old clothes to new products.