AsianFin – JD.com’s 2025 “618 Shopping Festival” wrapped up with record-breaking momentum, driven by surging user participation, AI-infused retail tools, and a deep push into intelligent services across sectors.
The e-commerce giant reported that as of 11:59 p.m. on June 18, more than 2.2 billion orders were placed across JD Retail’s online and offline platforms—an all-time high. The number of users placing orders more than doubled year-on-year, while sales from procurement and livestreaming surged 285%.
June 18 marked the peak of the festival, with JD’s headquarters emblazoned with the slogan “Fight to be No.1.” The company highlighted AI as a core driver of this year’s success, citing explosive growth in smart hardware sales and AI user engagement.
AI-related product categories posted stellar results:
AI-powered laptops: up 151% YoY
AI tablets: up 149%
AI learning devices: up 120%
Searches for Chinese-brand AI products rose more than 120%, with AI smartphones, glasses, and PCs topping consumer wish lists.
JD’s large AI models—integrated throughout its retail, logistics, and healthcare businesses—processed 1.3 billion calls during the event. Over 14,000 AI agents were in operation internally, while more than 17,000 brands leveraged JD’s digital human livestreaming services. The Jingyan AI shopping assistant handled over 50 million user interactions, while JD Health’s AI capabilities served 30 million users.
The company’s intelligent retail infrastructure is now largely AI-native. JD executives revealed plans to expand AI-generated digital humans to serve over 50,000 brands by the end of 2025, cutting costs and boosting operational agility. More than 90% of marketing copy is now generated by AI, significantly reducing production cycles and enabling real-time personalization.
“Our digital humans today operate at one-tenth the cost of human hosts and outperform over 80% of them in livestream conversions,” one JD executive said. JD’s Jingmai Smart Merchant Assistant has boosted decision accuracy to 90%, reduced manual inquiries by 16%, and saved over 1,500 hours weekly.
Beyond retail, JD’s large models are reshaping the entire supply chain:
Logistics: AI assisted 500,000+ frontline workers, automating routing and fulfillment
Healthcare: JD Health’s AI served over 50 million users, while its hospital AI system powered care for 1.8 million patients
Industrial: JD Industrial launched JoyIndustrial, cutting technical and compliance costs by over 30%
Enterprise AI: JD’s JoyAgent 2.0 now powers smart customer service across enterprise verticals, with a 200% MoM usage increase over the past quarter
JD is also betting big on embodied AI—the fusion of large models with physical robots. The company’s Joy Insideplatform debuted at 618, showcasing humanoid and quadruped robots from partners like SenseTime and Ezviz, all equipped with JD’s proprietary LLMs.
“This is the next phase of AGI—where intelligence transitions from code to the physical world,” a JD executive told Bloomberg. “We believe multi-modality, physical execution, and deep reasoning are the key pathways to embodied intelligence.”
JD now operates multiple proprietary foundation models:
3B, 10B, and 81B Pro base models
81B-T1 deep reasoning model
470B Preview, benchmarked against OpenAI’s GPT-4o
750B MoE (40B active), delivering GPT-4-class performance with improved efficiency
Its AIGC platform Jingdiandian handles over 10 million content requests per day, supporting 800,000+ merchantsand 100 million+ products. Meanwhile, Jingzhuntong ad tools and Jingxiaozhi 5.0 customer service suite are driving higher sales and satisfaction—with manual intervention rates now as low as 28% and customer satisfaction reaching 71%.
JD’s AI-powered 618 Festival highlights its broader shift from traditional e-commerce to an intelligent retail and services platform. With AI deeply embedded across sales, logistics, healthcare, and industrial operations, JD is now positioning itself as not just a marketplace—but a full-stack AI commerce infrastructure provider.
As competition heats up in the generative AI race, JD’s early investment in model development, commercialization, and hardware integration is giving it a unique edge in turning large models into large-scale business outcomes.