JD.com Launches Oxygen AIIC for Scalable E-commerce Item Understanding.

Oxygen AIIC, Chan Long, Chao Liu, Chaofan Chen, Chaohui Dong, Chunyuan Guo, Danping Liu, Debin Liu, Deping Xiang, Fulai Xu, Guangyue Liu, Hao Li, Huichun Hu, Jian Yang, Jianan Wang, Jianbo Zhao, Jiaoyang Li, Jiaxing Wang, Jinglong Li, Jinjin Guo, Jun Fang, Jun Liu, Kai Zhou, Li Wang, Lili Gao, Liying Chen, Luning Yang, Mengdi Zhou, Pengzhang Liu, Qi Lv, Qianyun Wang, Qixia Jiang, Ruyue Li, Shimu Liang, Shuxing Wang, Sijie Zhang, Siqi Li, Tianhao Gao, Wang Ke, Weihu Huang, Wencan Lai, Wenjie Zhang, Xiaohui Zhang, Xiaojing Dong, Ya Liu, Yifeng Zhang, Yixiang Wang, Yongtai Zhang, Yongyi Liao, Zhaoru Chen, Zhen Chen, Zhiyong Ma, Zhiyuan Liu, Zhongwei Liu, Ziyan Xing· June 29, 2026 View original

▶ The 2-minute explainer

Summary

JD.com has developed Oxygen AIIC, an industrial-scale platform leveraging LLMs and VLMs to enhance item understanding, management, and applications for its vast e-commerce catalog. This system addresses challenges like fast-emerging concepts, high-quality knowledge production for billions of SKUs, and diverse downstream requirements.

JD.com, a major global e-commerce platform, has unveiled the JD Oxygen AI Item Center (Oxygen AIIC), a sophisticated system designed to manage and understand its enormous catalog of products. Serving hundreds of millions of users and merchants with tens of billions of SKUs, the platform aims to improve customer experience, reduce operational costs, and boost efficiency through high-quality, structured item knowledge. Oxygen AIIC tackles key challenges such as rapidly evolving product concepts, the need for consistent knowledge across a massive inventory, and varied demands from different business functions. The Oxygen AIIC platform is built upon four foundational pillars. It employs human-AI collaboration for dynamic ontology engineering, supporting millions of entries. A "Semantic Search then Discrimination" (S2D) architecture, combined with throughput optimization, enables scalable production of an AI Item Library for billions of SKUs. The system also features self-evolving LLMs/VLMs for item understanding, achieving high precision and recall in knowledge production. Finally, a unified item tunnel acts as a central hub for data and services. This system is already deployed across core business areas like search and recommendations, demonstrating significant improvements in search traffic coverage, reduced item information quality issues, and increased automation in attribute filling.

Why it matters

This demonstrates a successful, large-scale application of LLMs/VLMs in e-commerce for critical operational efficiency and customer experience, offering a blueprint for similar challenges in other industries with vast inventories.

How to implement this in your domain

  1. 1Analyze current product information management systems for scalability and accuracy bottlenecks.
  2. 2Investigate the feasibility of integrating LLM/VLM technologies for automated item attribute extraction and categorization.
  3. 3Develop a strategy for human-AI collaboration in ontology management to handle evolving product concepts.
  4. 4Pilot an AI-driven item understanding system on a subset of products to measure improvements in data quality and operational efficiency.
  5. 5Establish metrics to track the impact on search relevance, recommendation accuracy, and manual intervention rates.

Who benefits

E-commerceRetailManufacturingLogisticsData Management

Key takeaways

  • JD.com's Oxygen AIIC uses LLMs/VLMs for industrial-scale item understanding in e-commerce.
  • It addresses challenges of concept emergence, knowledge quality, and diverse requirements.
  • The platform features human-AI ontology engineering and a scalable knowledge identification architecture.
  • Oxygen AIIC has delivered measurable gains in search coverage, data quality, and automation.

Original post by Oxygen AIIC, Chan Long, Chao Liu, Chaofan Chen, Chaohui Dong, Chunyuan Guo, Danping Liu, Debin Liu, Deping Xiang, Fulai Xu, Guangyue Liu, Hao Li, Huichun Hu, Jian Yang, Jianan Wang, Jianbo Zhao, Jiaoyang Li, Jiaxing Wang, Jinglong Li, Jinjin Guo, Jun Fang, Jun Liu, Kai Zhou, Li Wang, Lili Gao, Liying Chen, Luning Yang, Mengdi Zhou, Pengzhang Liu, Qi Lv, Qianyun Wang, Qixia Jiang, Ruyue Li, Shimu Liang, Shuxing Wang, Sijie Zhang, Siqi Li, Tianhao Gao, Wang Ke, Weihu Huang, Wencan Lai, Wenjie Zhang, Xiaohui Zhang, Xiaojing Dong, Ya Liu, Yifeng Zhang, Yixiang Wang, Yongtai Zhang, Yongyi Liao, Zhaoru Chen, Zhen Chen, Zhiyong Ma, Zhiyuan Liu, Zhongwei Liu, Ziyan Xing

"arXiv:2606.28070v1 Announce Type: new Abstract: JD.com, one of the world's largest e-commerce platforms, serves over 700 million active users and millions of merchants, with a catalog of tens of billions of SKUs. At this scale, high-quality, structured item knowledge underpins a…"

View on X

Originally posted by Oxygen AIIC, Chan Long, Chao Liu, Chaofan Chen, Chaohui Dong, Chunyuan Guo, Danping Liu, Debin Liu, Deping Xiang, Fulai Xu, Guangyue Liu, Hao Li, Huichun Hu, Jian Yang, Jianan Wang, Jianbo Zhao, Jiaoyang Li, Jiaxing Wang, Jinglong Li, Jinjin Guo, Jun Fang, Jun Liu, Kai Zhou, Li Wang, Lili Gao, Liying Chen, Luning Yang, Mengdi Zhou, Pengzhang Liu, Qi Lv, Qianyun Wang, Qixia Jiang, Ruyue Li, Shimu Liang, Shuxing Wang, Sijie Zhang, Siqi Li, Tianhao Gao, Wang Ke, Weihu Huang, Wencan Lai, Wenjie Zhang, Xiaohui Zhang, Xiaojing Dong, Ya Liu, Yifeng Zhang, Yixiang Wang, Yongtai Zhang, Yongyi Liao, Zhaoru Chen, Zhen Chen, Zhiyong Ma, Zhiyuan Liu, Zhongwei Liu, Ziyan Xing on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses