Japanese retailer Cainz expands daily plastic waste collection with AI

Cainz Corporation, a popular home center chain with stores across Japan, has moved on to Phase 3 of its “Pla-Recycle” resource circulation project in Kakegawa City, Shizuoka Prefecture, together with their partners. They are introducing an automated collection system powered by artificial intelligence. The three-month demonstration will run until 13 December 2026 across the Cainz Kakegawa and Cainz Daito stores.

Cainz’s framework uses a LINE mini-application equipped with AI image recognition technology. Before visiting a store, residents upload photographs of their discarded plastic items to verify eligibility. Pre-registered users can then unlock dedicated collection sheds on site during operational hours between 9:30am and 6:00pm daily. Cameras inside the storage units use image recognition to monitor fill levels remotely, allowing facility managers to schedule transport without requiring store staff to conduct manual checks.

By automating sorting and access, the project has expanded the scope of accepted materials from nine item types in 2025 to approximately 50 categories of large household plastics, whilst increasing collection frequency from scheduled days to a daily service.

Cainz began the Pla-Recycle partnership in 2024 to support Kakegawa City’s strategy to reach carbon neutrality by 2050, positioning its retail stores as local circular economy hubs termed “Circular Stations”. Phase 1 collected approximately 65 tonnes of product plastics to test municipal feasibility, while Phase 2 focused on quality control through manned reception points. To demonstrate closed-loop material flow, a portion of the recovered plastics from the initial trial was processed into recycled resin and manufactured into shopping baskets. These baskets went on sale across Cainz stores in Shizuoka Prefecture in October 2025.

Cainz aims to lower operational costs and make household plastic recycling more accessible for local residents by combining mobile application pre-sorting with unmanned drop-off points.