CJ Logistics Doubles Down on “Physical AI,” Backs Startup RLWRLD to Build Robot Brains

South Korean supply chain giant CJ Logistics is moving to solve the most difficult part of the humanoid equation: the brain. On November 23, the company announced a strategic memorandum of understanding (MOU) and an equity investment in RLWRLD, a startup specializing in "Physical AI" and foundational robotics models.

The move signals a shift in CJ Logistics’ strategy from simply acquiring hardware to actively cultivating the software ecosystem needed to make autonomous warehouse robots a reality. The company formalized this pivot just days later at a National Assembly seminar, declaring "Physical AI" as the core engine for its future operations.

The Shift to "Logistics AX"

Speaking at a seminar on November 24 regarding AI innovation in the logistics industry, CJ Logistics presented its strategy for "Logistics AX" (AI Transformation). Unlike traditional Digital Transformation (DX), which digitizes processes, AX aims to deploy AI that can physically perceive, judge, and act within the real world.

"Physical AI is the core technology that will establish an autonomous operating system for logistics centers," said Goo Sung-yong, head of CJ Logistics’ TES Automation Development, during the seminar.

He argued that logistics centers are the optimal testbed for this technology because they generate massive amounts of data daily while handling items of infinite shapes and sizes. The company believes successful deployment here could cascade into other industries like manufacturing and retail.

The Missing Piece: The "Brain"

CJ Logistics has spent much of 2025 assembling a roster of domestic hardware partners to build its "AI Alliance".

• Hardware Platforms: The company previously partnered with Rainbow Robotics and Robotis to secure mobile manipulator platforms.
• Sensors & Actuators: It is collaborating with Aidin Robotics for robot hand technology.

However, hardware without adaptive intelligence is of limited use in the chaotic environment of a distribution center. Under the new agreement with RLWRLD, CJ Logistics will participate in the startup’s Seed2 funding round and collaborate on developing a Robotics Foundation Model (RFM).

"The competitiveness of future logistics will depend not on simple equipment, but on how advanced an AI brain a company possesses," said Jonghoon Lee, Head of Management Support at CJ Logistics.

Focusing on Fine Motor Skills

While walking is a largely solved problem for modern humanoids, dexterous manipulation—picking up a crushed box, handling a fragile package, or sorting varied items—remains a massive hurdle.

The collaboration focuses specifically on this bottleneck. CJ Logistics plans to identify specific processes in its centers ripe for automation and provide operational data regarding picking, sorting, and packaging. In exchange, RLWRLD will use this real-world data to train its "Physical AI" models to control robotic hands with high precision.

RLWRLD, led by CEO Jung-hee Ryu (formerly of venture capital firm FuturePlay), specializes in hardware-agnostic robotics models. The startup utilizes "4D+ multimodal data" to create brains that can adapt to different robot bodies, a capability recently recognized when it was selected for Amazon Web Services' (AWS) Generative AI Accelerator Cohort 3.

A National Effort

The push for "Physical AI" is being coordinated closely with South Korea's broader industrial policy. The seminar where these strategies were unveiled was hosted by members of the National Assembly and sponsored by the Humanoid M.AX Alliance.

By integrating Rainbow Robotics’ mobile platforms, Aidin’s sensor-rich hands, and now RLWRLD’s cognitive models, CJ Logistics is attempting to position itself as a systems integrator for South Korea's robotics sector. If successful, the project could serve as a proof-of-concept for the government's goal to achieve mass production of humanoid robots by 2029.