JD.com's Robot Empire: 3 Million Robots and 700,000 Couriers Retrained

At the JD.com Global Technology Exploration Conference (JDD) held on September 9, 2026, JD Logistics showcased its full "Wolf Pack" robotics lineup for the first time, including nine models spanning warehousing, sorting, transport, and last-mile delivery. Robots such as the "Smart Wolf," "Ground Wolf," "Flying Wolf," "Warehouse Wolf," and "Mother Wolf" were presented. The company also unveiled its "Physical AI Acceleration Plan": within five years, it will procure 3 million robots, 1 million autonomous vehicles, and 100,000 drones. This is the most aggressive unmanned logistics timetable ever announced by a Chinese logistics enterprise.

The plan encompasses several parallel dimensions: on the data front, JD Cloud aims to collect over 10 million hours of real-world human activity video in two years to train physical AI models. On infrastructure, JD.com will build 80 RoboBase robot industrial parks across the nation in five years, creating the world's largest robot maintenance and repair network; eight regional repair centers are already operational. On procurement, the company will deploy 3 million robots, 1 million autonomous vehicles, and 100,000 drones. Newly unveiled products include the "Mother Wolf" for 24/7 unmanned pharmacy operations, a low-temperature version of the "Smart Wolf" for cold-chain logistics, the embodied picking robot "Warehouse Wolf," the L4-level "Lone Wolf" sixth-generation delivery vehicle Plus, and the "Flying Wolf" L05—the industry's first logistics drone supporting fully unmanned operations.

To address the impact on its 700,000 couriers, JD.com proposed the "Nirvana Plan." Richard Liu, the company's founder, had earlier stated at an APEC forum that future deliveries would not require couriers, but he did not want his 700,000 employees to lose their livelihoods. Consequently, JD.com has signed agreements with 120 vocational and technical schools, sending frontline blue-collar workers through systematic training in robot repair, maintenance, and fault diagnosis. The logic is clear: robots replace delivery personnel at the last mile, but robots themselves need maintenance, a structural gap requiring human workers. By August 2026, the first batch of frontline employees had completed the retraining and formally took up positions as robot maintenance engineers.

This initiative addresses a severe gap in the after-sales service system of the entire robot industry. Data shows that global humanoid robot shipments reached approximately 13,000 to 16,000 units in 2025, with China accounting for about 90%. Morgan Stanley predicts that Chinese humanoid robot sales will double to about 28,000 units in 2026. However, many faulty robots currently must be returned to factories for repairs, a time-consuming and costly process.

JD.com's differentiation strategy is to avoid participating in the profit distribution of robot bodies themselves. Instead, it transfers its supply chain management, doorstep service network, and distribution capabilities accumulated in the consumer electronics sector to the robot industry, attempting to become the underlying infrastructure behind all robot brands—"both a hub connecting manufacturing and applications and a guardian ensuring full-lifecycle operation." This positioning leverages JD.com's strengths without competing directly with robot makers.

Yet transforming 700,000 people from delivery to robot repair is not a simple slogan. A workforce of 700,000 is equivalent to a medium-sized city's population, with varying educational backgrounds, heavy family burdens, and limited study time. Robot maintenance engineers require composite knowledge of circuits, sensors, software systems, and fault diagnosis—essentially different from changing tires or tightening screws. Commentators have noted that robot maintenance is one branch but should not be the only exit for all 700,000 people; transformation should pave different roads for different individuals. JD.com's "Physical AI Acceleration Plan" is fundamentally a systematic experiment binding technological substitution with employment transformation. The acceleration of unmanned operations is no longer just a matter of efficiency; it raises a key question: when machines replace human roles, what should people do?