Kioxia Corporation has launched an AI image recognition technology. This tool automatically identifies products in logistics workflows. It was developed with Tsubakimoto Chain Co. and EAGLYS Inc. This innovative system boosts automation and improves efficiency. It helps logistics organizations adapt to market demands. Plus, it optimizes costs and keeps service quality high. KIOXIA AiSAQ™ and the company’s Memory-Centric AI technologies are key to this progress. They make it easier to deploy AI as product categories grow and change. The jointly developed solution will be showcased at the 2025 International Robot Exhibition.
Logistics providers face several challenges. Shipment volumes are rising. Product assortments are becoming more complex. Also, labor shortages continue due to the growth of e-commerce. Traditional image recognition systems usually rely on deep-learning models. These models need fine-tuning or retraining when new seasonal or variant products come into the supply chain. Ongoing maintenance can slow down operations. It also raises power use and costs, especially for facilities with large inventories.
KIOXIA AiSAQ software, leveraging the power of Memory-Centric AI, offers a much more flexible and scalable solution. By storing large product data, including images, labels, and other important attributes, directly in high-capacity storage, the system can rapidly add new items without having to retrain the base model. To keep speed and efficiency with growing data, the technology indexes info in memory. It then transfers this indexed data to SSD storage, which speeds up searches and retrieval times.
Also Read: BAE Systems receives $390 million to produce Bradley A4s
Kioxia, Tsubakimoto Chain, and EAGLYS are teaming up for a new innovation. You can see it at the 2025 International Robot Exhibition. The event will take place from December 3 to 6 at Tokyo Big Sight, specifically at the Tsubakimoto Chain Booth E6-23. The exhibition is a top global event for robotics and automation. It will feature the image recognition system. During the demonstration, products moving down a conveyor will be photographed, analyzed, and accurately sorted by referring to stored metadata, with the intent of illustrating how logistics sites can improve accuracy, responsiveness, and workflow efficiency even as product mix complexity continues to rise.