RLWRLD, the South Korean physical AI startup, demonstrated two robots performing packing and sorting tasks through real-time coordination with no human in the loop using its RLDX-1 foundation model

RLWRLD just showed two robots packing and sorting together, coordinated in real time with no human in the loop. 🤖 The system is powered by RLDX-1, its dexterity-first foundation model built for robot hands. The interesting part is not just the box-packing demo. RLDX-1 is designed around the hard part of real-world manipulation: vision, language, motion, memory, tactile signals, torque, and action generation running inside one control stack. That matters for warehouse and factory work. A robot does not only need to “see” the object. It has to remember the task context, feel contact, adjust grip, and finish the motion without dropping the workflow. RLWRLD is a South Korean physical AI company founded by Jung-Hee Ryu. The company raised about $26 million in Seed 2 funding, bringing total seed funding to about $41 million, with strategic investors across logistics, retail, manufacturing, telecom, and finance. Their bet is clear: industrial robots do not need better demos first. They need hands that can execute messy, contact-rich work at production speed.