跳转到主要内容

JD.com Launches Physical AI Acceleration Plan, Aiming to Become the World's Largest Physical World Operations Center

At the recent JD.com Global Technology Explorer Conference, Cao Peng, Chairman of JD.com's Technical Committee and President of JD Cloud, gave people new expectations for the implementation of AI. He said that physical AI is striding towards large-scale application, and the focus of competition has shifted from single-point competition between models and entities to systematic competition in computing power, data, and scenarios.

Cao Peng emphasized that JD.com's approach to AI is not just停留在论文里, but is solidly implemented in production lines, warehouses, distribution stations, and every link of the entire supply chain. In his words, it is 'run' out order by order. Based on this philosophy, JD.com officially launched the 'Physical AI Acceleration Plan', aiming to continuously invest in six key areas: data, machinery, manufacturing, sales, logistics, and services, paving the way for the development of embodied intelligent robots.

Six Goals: From 10 Million Hours of Data to 300,000 Robots

First, let's look at data. JD.com plans to build the 'world's largest embodied data training center', intending to collect over 10 million hours of real human scenario data within two years, while integrating network data, simulation data, and real robot data to train an embodied intelligent large model that can 'understand-generate-act'. This sounds grand, but data is the fuel for AI; without sufficient data, even the smartest algorithms cannot run.

Next, manufacturing. JD.com's goal is to establish the 'world's largest robot manufacturing base', planning to layout 80 bases nationwide in the next five years, providing full lifecycle services from display, delivery, maintenance to R&D, design, assembly, data collection, production, and iteration. On the manufacturing side, they aim to become the 'world's largest supplier of core robot components', investing in core components and building their own factories in the next three years, helping 100 complete machine manufacturers reduce costs and increase efficiency, while promoting 100 component suppliers to double their performance.

In terms of sales, by 2028, JD.com plans to invest over 10 billion yuan to help 100 brands achieve independent sales exceeding 1 billion yuan, bringing robots into thousands of terminal scenarios. In the logistics field, JD.com has already achieved full-chain joint development in warehousing, sorting, transportation, and delivery. In the next five years, they plan to procure over 300,000 robots, 100,000 unmanned vehicles, and 100,000 drones. For services, JD.com will establish after-sales and service networks in over 100 countries globally, training more than 100,000 robot after-sales service engineers.

Cao Peng stated that as these goals are gradually implemented, JD.com will promote AI from thousands of industries to global households, build the world's largest physical world collaboration system, and ultimately become a leading global physical world operations center.

Key Points

  • JD.com releases 'Physical AI Acceleration Plan', focusing on six areas: data, machinery, manufacturing, sales, logistics, and services.
  • Plans to collect over 10 million hours of real scenario data within two years and build 80 robot manufacturing bases.
  • Procure 300,000 robots, 100,000 unmanned vehicles, and 100,000 drones in the next five years.
  • Goal to become the world's largest physical world operations center.