This cooperation pursues both scenario‑based implementation and cutting‑edge innovation. On one hand, it will foster integration between industrial large models and high‑performance edge‑side computing power, developing scalable products and solutions for use‑cases including machine vision, equipment maintenance and operational safety. On the other hand, both parties will jointly explore forward‑looking technologies such as industrial world models, advancing industrial AI from perception and recognition toward comprehension, deduction and prediction, and building replicable and scalable best‑practice templates for industrial intelligence.
Wang Peinuan, CEO of JD Industrial, commented that through joint optimization with AMD, JD Industrial will substantially strengthen its industrial‑grade AI technical service capabilities and real‑world application value. “We will keep leveraging our integrated strengths of precise professional data and deep industry‑specific scenarios. Adhering to JD Industrial’s innovation‑driven and value‑oriented mindset, we will deploy AI into every real‑world, niche scenario, enabling innovative applications to deliver tangible cost reduction, efficiency gains and performance improvements for industries.”
Pan Xiaoming, Senior Vice President and President of AMD Greater China, stated that this collaboration represents an innovative leading paradigm for industrial intelligence jointly explored by both sides, enabling AI to evolve beyond “detection and recognition” toward “comprehension, deduction and prediction” of real‑world production environments. “We aim to propel industry advancement from automation and digitalization to intelligence featuring continuous perception, understanding and learning, empowering every piece of equipment, every production procedure and every manufacturing decision with AI.”
Deepening Integration of Models and Computing Power to Expand Industrial Edge‑AI Applications
JD Industrial is a leading AI‑powered industrial supply‑chain technology and service provider. Driven by AI innovation, the company has built a digital‑intelligent highway connecting supply and demand, spanning diverse industries and compatible with multiple material categories. By “boosting data flow while minimizing physical goods movement”, it enables efficient collaboration across all industrial‑chain links and significantly improves supply‑chain resilience and cost‑effectiveness. As a global leader in high‑performance and adaptive computing, AMD boasts profound technical expertise and a world‑class product portfolio covering CPUs, GPUs, FPGAs, adaptive SoCs and deep‑learning inference accelerators. Its robust computing foundation fuels intelligent transformation across countless industries. AMD empowers creative ideas to translate into tangible outcomes via its reliable, open and flexible computing infrastructure.
The partnership will fully combine JD Industrial’s comprehensive strengths in industrial scenarios, industrial large models, supply‑chain networks, channel reach and customer resources with AMD’s leading advantages in high‑performance computing chips, hardware‑software co‑optimization and global technical ecosystem. Both parties will jointly advance industrial AI innovation and real‑world deployment.
Under the agreement, JD Industrial’s industrial large models will undergo adaptation and co‑optimization with AMD Ryzen AI Max series processors. Targeting industrial terminals and edge‑computing environments such as factories and workshops, the two sides will co‑develop AI terminal products and applications for on‑site industrial operations. Joint solutions will be defined, developed, validated and deployed as benchmark projects across key industrial scenarios including smart manufacturing, predictive equipment maintenance, industrial visual inspection, safety inspection, intelligent production scheduling and supply‑chain collaboration. This will realize integrated deployment and large‑scale commercialization of industrial large models on edge terminals.
Meanwhile, both parties will deliver AI terminal products, application solutions and services for multinational corporations, international enterprises and Chinese enterprises expanding overseas. Leveraging JD Industrial’s diversified sales channel network, the collaboration will comprehensively drive sales growth and market penetration for AMD‑branded products.
Advancing Industrial World Models: Exploring the Transition from On‑site Perception to Proactive Decision‑making
While rolling out industrial edge‑AI applications, the partners will also conduct cutting‑edge research on industrial world models. Drawing on JD Industrial’s accumulated industrial data, domain expertise and operational workflows, vertical industrial world models will be developed. Combined with AMD’s high‑performance computing and edge‑deployment technologies, the collaboration will explore industrial‑intelligence applications capable of on‑site condition understanding, dynamic deduction and risk forecasting. The two parties strive to create a benchmark case for deep integration between industrial world models and high‑performance edge‑side computing power, offering replicable and scalable reference implementations for industrial AI adoption.
This framework further extends industrial AI applications from centralized analytics to real‑time edge‑local decision‑making. Complex on‑site risk assessment, trend deduction, task interpretation and decision support can be executed locally, cutting response latency, enhancing data security and boosting deployment flexibility.
The edge‑side AI innovations delivered by the joint effort will bring substantial operational upgrades for real‑world industrial settings. For instance, personnel may inadvertently enter vehicle travel paths inside production workshops or warehousing zones, creating safety hazards. By combining industrial world models with edge‑side computing power, future states of personnel, equipment, materials, environments and operational workflows can be simulated. Factoring in site spatial layouts, vehicle trajectories, human‑behavior patterns and visual occlusions, potential collision risks can be identified several seconds in advance. The system can trigger audio‑visual alerts, send evasion prompts to forklift operators or command vehicles to decelerate proactively.
(Reprinted from https://news.eccn.com/）