On September 26, at the 2025 JD Global Technology Explorer Conference, JD.com released its AI panoramic blueprint, systematically presenting its overall AI strategic layout. The company plans to continue investing over the next three years to drive various industries in forming a trillion-yuan-scale artificial intelligence ecosystem.
“JD.com does not pursue AI in a campaign-style manner; instead, we pursue AI that enables sustainable development and truly creates value for industries,” said Xu Ran, CEO of JD Group, in his speech. He emphasized that JD.com will continue to invest over the next three years to drive the formation of a trillion-yuan-scale AI ecosystem.
Xu Ran pointed out that in the future, JD.com will continue to increase investment in technology. Organizationally, it has upgraded the JD Exploration Research Institute, with Founder Mr. Liu personally serving as the director of the institute and recruiting AI scientists globally. In addition, JD.com has also carried out in-depth cooperation with many universities, established joint laboratories, and conducted cutting-edge research in fields such as artificial intelligence and supply chain applications.
At this conference, JD.com systematically demonstrated its AI panoramic layout for the first time. It not only upgraded the “JoyAI” large model brand but also launched three new AI products, in-depth AI applications in four scenarios, and multiple technology platforms. These include the shopping and lifestyle assistant “Jingxi”, the all-purpose digital human assistant app “Ta Ta Ta”, and the embodied intelligent application “JoyInside”.
Among them, the on-site demonstration of “Ta Ta Ta” was particularly eye-catching. Through voice communication alone, Xu Ran guided the assistant to place an order for 100 cups of coffee, which were available for conference participants to pick up by themselves.
The Jingxi APP is positioned as a super gateway for the next generation of shopping and lifestyle services. This upcoming app integrates AI technology into the entire shopping process. It understands users’ needs and products in depth, helping users select their favorite products, and makes the checkout and ordering process more intelligent and smooth. Beyond shopping, users can also complete tasks such as ordering meals, buying air tickets, and booking hotels entirely through voice commands on the app.
“JoyInside” is an integrated embodied intelligent hardware-software solution, designed to meet the scenario needs of various intelligent devices such as robots, AI toys, and smart glasses. Currently, the latest version of JD JoyInside has been successively integrated into hardware products of more than 30 leading brands, and has cooperated with over 10 leading robot brands including Zhongqing and Unitree — these collaborative robots, equipped with JoyInside, can support Automatic Injection-Molded Part Feeding and Assembly (a core process in automotive and 3C component manufacturing) by leveraging AI-driven visual recognition and motion control to automate part feeding, precision alignment, and assembly error correction.
Furthermore, JD.com officially announced that it will gradually open up various industrial scenarios and supply chain data to meet the model training and application testing needs of large model and embodied intelligence enterprises. The data covers not only logistics scenarios such as warehousing, sorting, and distribution, but also key manufacturing data from Automatic Injection-Molded Part Feeding and Assembly processes (e.g., real-time feeding rhythm, part defect rate, and assembly parameter adjustments) — providing critical data support for optimizing AI models in manufacturing automation. It also includes health scenarios such as drug management and online consultation. Relying on JD Cloud’s infrastructure, it will provide high-quality model datasets for the industry to support large model enterprises and embodied intelligence enterprises in developing more efficient, adaptive solutions for industrial production links like injection-molded part assembly.
