Company profile
RLWRLD
RLWRLD develops robotics foundation models for dexterous manipulation. Its RLDX-1 model is built around five-finger hand control and complex real-world tas...
Facts
- Segment
- FOUNDATION_MODELS
- Country
- South Korea
- Headquarters
- Seoul
- Founded
- 2024
- Website
- https://www.rlwrld.ai/
About RLWRLD
RLDX-1 is a dexterity-first foundation model built around five-finger manipulation rather than gripper-only motion. Its multimodal policy processes vision and language together with force, tactile contact and working-memory signals, allowing robot hands to coordinate contact timing and torque during precise industrial tasks.
RLWRLD develops the model from task data collected with industry partners and is building shared dexterity infrastructure around it. The company introduced RLDX-1 in 2026 alongside DexBench and humanoid-data standards intended to compare manipulation performance across hands and robot embodiments. Operations span the United States, South Korea and Japan, with the current programme centred on model and data development rather than a proprietary humanoid body.
Market context
DEVELOPMENT
Summary
Type: AI platform
Core: RLDX-1
Function: Manipulation
Audience: Robot OEMs