EXPLORINEER
en

JP — ITmedia

AEON Pushes Logistics Overhaul With Physical AI, Reaches 50% Automation at Distribution Hub Amid Labor Shortage

Robotics2 min read

Japanese retailer AEON and robot software maker Mujin opened a delivery center that uses "physical AI" robots. Half of its work is now automated, and AEON plans to reach 80% by 2030.

A warehouse run by AI-powered robots

On October 29, AEON and Mujin showed reporters a delivery center called AEON Kawaguchi Higashi XD in Kawaguchi, Saitama. The center started full operation that day. Its robots use physical AI, which means artificial intelligence that helps machines move and act on their own. AEON plans to bring this technology to its warehouses across Japan.

Key numbers

  • Automation rate at the center: about 50% of all work, with a goal of 80% by 2030
  • More than 90 robots are running, and they save the work of 39 people per day
  • 82 AGVs (automated guided vehicles, or driverless carts) move loose items to temporary storage
  • The center supplies about 200 small AEON Maibasket stores in the Tokyo area
  • A robot maker's boss says robots work 3 to 4 times as fast as people

Why stores need this help

Japan has fewer working-age people every year. Jobs in warehouses and stores are hard to fill, so companies want machines to do the work. A research firm estimates that the global physical AI market will grow from about 5.0 billion dollars in 2025 to about 82.8 billion dollars in 2035. The Japanese government also lists physical AI as one of its 17 key strategy areas.

Why it matters

This shows that Japanese companies can use robots to solve real labor shortages. If Japan gets good at physical AI, its businesses may compete better around the world. Experts warn that robots must be safe so they do not harm people. A top researcher says the full version will take many years to build, so companies should start with what they can do now.

What to watch

  • Whether AEON reaches its goal of 80% automation by 2030
  • Whether the robots can also handle the full process, from receiving goods to shipping them
  • How AEON applies the system to its other warehouses and store deliveries
  • How the government's plan to invest about 10.5 trillion yen by fiscal 2040 is used

Glossary

Mujin
A Japanese company that makes software to control robots.

Same story, different eyes

Ad

Add 150 words. Learn 10 a day.

Dicter word books send only the words you learn to your review notes.

More in this topic

RoboticsZDNet Korea

Humanoid Robot Climbs Ladders and Balances After Single Training Session

Researchers from ETH Zurich and collaborators developed a humanoid robot control technology that, after one round of training, lets a Unitree G1 climb ladders, jump obstacles, and recover its balance after falling. The system, called Attention-based Contact Planning (ACP), plans which body parts, such as hands, feet, knees, and elbows, should touch the environment, and a separate controller adjusts the joints to reach those points. In simulation tests across eight complex obstacle courses, ACP achieved an average 92.5% success rate, compared with 58.8% for a baseline method, though the authors note that real-world success rates were not established by this figure.

Roboticsheise online

heise+ | Machine Space Management: Build Your Own Adapter Plate System

The article describes a DIY adapter plate system for workshops. Each machine is permanently mounted on a wooden plate with a standardized hole pattern, which can be bolted to any workbench, so the workshop can be reconfigured as needed. The article is a paywalled heise+ piece; only an excerpt is available.

RoboticsTechNode

WeRide’s physical AI data subsidiary raises tens of millions of dollars

Jingshuo, a WeRide subsidiary that builds data infrastructure for physical AI, has raised tens of millions of dollars in pre-Series A funding from investors including Vision Knight Capital, Hong Kong Science and Technology Parks Corporation, and Sirius Capital. The company plans to use the money to grow its R&D team, develop its data platform, world models, and reusable robot skill systems, and produce more high-quality datasets. It remains a majority-owned WeRide subsidiary and is still consolidated into the parent's financials.

RoboticsThe Register

Please state the nature of the medical emergency: Google is building a robot medic for NASA

Google is working with NASA to develop a robotic medical assistant that could help astronauts on future deep-space missions. The robot builds on CMO-DA, an AI medical decision-support system being evaluated on Earth-based copies of ISS computers, and is designed to handle cargo, logistics, and routine maintenance while assisting a crew medical officer with medical tasks. The project remains at the proof-of-concept stage, with early lab tests reporting an initial 85 percent success rate in locating and moving objects.