RLWRLD
Category: Robotics / Embodied AI
A Physical AI startup developing Robotics Foundation Models (RFM) that enable robots to achieve human-level dexterity and cognitive abilities through five-fingered hand manipulation and real-world sensor perception. RLWRLD was founded in 2024. The company is led by Jung-hee Ryu. Based in San Francisco, United States (HQ after corporate flip), with offices in Seoul, South Korea and Tokyo, Japan. Team size: N/A. Total funding raised: $41.0M. Latest round: Seed 2. Key investors include Hashed, FutureAsset Venture Investment, Global Brain, LG Electronics, SK Telecom, DRB Dongil, ANA Holdings, KDDI, Mitsubishi Chemical, Shimadzu, Microsoft, NVIDIA, AWS, Headline Asia, Z Venture Capital, CJ Logistics, Lotte Ventures, Hanwha Asset Management, Mirae Asset - E-mart Investment Fund I, Hyosung Ventures, Smilegate Investment, T Investment.
- Founded
- 2024
- Headquarters
- San Francisco, United States (HQ after corporate flip), with offices in Seoul, South Korea and Tokyo, Japan
- Team size
- N/A
- Total funding
- $41.0M
Value proposition
Enables robots to see, think, and use their hands like humans by training foundation models on real-world sensor data and industrial workflows, addressing labor crises in manufacturing and logistics while achieving five-finger dexterity capabilities not yet demonstrated by competitors like Tesla, Figure AI, and 1X.
Products and solutions
RLDX-1 (Robotics Foundation Model) - dexterity-first foundation model for robot hands, Five-finger robot hand manipulation system for humanoid robots, Real-world sensor data processing and cognition platform, Vision-Language Model (VLM) integration for robot control optimization, DexBench - benchmark for dexterity evaluation, Next-generation reference humanoid robots (co-developed with We Robotics)
Unique value
Focuses exclusively on Physical AI models trained on real-world sensor data, robotic systems, and industrial workflows from East Asia's top manufacturing ecosystems—distinct from text/code/image-based generative AI competitors. Claims breakthrough five-finger dexterity capabilities that competitors haven't publicly demonstrated.
Target customer
Large-scale industrial enterprises with manufacturing and logistics operations, humanoid robot developers, and retail companies seeking automation solutions in Korea, Japan, and India markets.
Industries served
Manufacturing, Logistics & Distribution, Retail, Robotics, Physical AI
Technology advantage
Combines embodied AI architectures with deep access to Korean and Japanese manufacturing data pipelines; leverages strategic investor relationships for immediate real-world testing environments; integrates proprietary RFM with global computing infrastructure (Microsoft, NVIDIA, AWS) to bridge the gap between LLM outputs and physical robot actions.
How they differentiate
Five-finger robot hand manipulation capability not yet publicly demonstrated by competitors; trains Robotics Foundation Models (RFM) directly in live industrial environments using real-world sensor data from East Asian manufacturing ecosystems; proprietary 'real-world data moat' from strategic investor partnerships enabling training in actual production settings
Main competitors
Tesla (Optimus), Figure AI, 1X Technologies
Key partnerships
Microsoft - cloud infrastructure, PoC projects in Asia's manufacturing/retail/hotel industries, NVIDIA - Isaac platform integration for robot hand manipulation learning, AI infrastructure support, AWS - cloud computing infrastructure support, We Robotics - strategic alliance for next-generation reference humanoid robot development, CJ Logistics - co-developing RFM for distribution centers, joint fulfillment center demonstrations, Lotte - retail space automation demonstrations, KDDI & Lawson - retail robot innovation projects targeting 30% operational efficiency improvement by 2030, LG Electronics - manufacturing data collaboration and industrial workflow integration, SK Telecom - telecommunications and manufacturing ecosystem partnerships, Upstage - Dokparmo consortium membership for AI model multimodal expansion, Robotics hardware partners: Robotis, Wonik Robotics, Rainbow Robotics, Sensor and AI partners: SOSLAB, Aidin Robotics, Vitsensing, Deepings, Flying, Seoul Robotics
Notable customers
CJ Logistics, Lotte, SK Telecom, KDDI
Major milestones
Emerging from stealth mode with $14.8M Seed 1 round in April 2025, Selected for 2025 AWS AI Accelerator program (October 2025), Won 1st place in Foundation Models category at Nebius Robotics & Physical AI Awards, receiving $150,000 in compute credits (December 2025), Raised $26M Seed 2 to expand strategic investor coalition (February 2026), Corporate relocation to United States announced (February 2026), Appointed Carl Choi as President of RLWRLD USA to lead US market development (March 2026), Released RLDX-1 dexterity-first foundation model, achieving SOTA on 8 global benchmarks (April/May 2026)
Growth metrics
Planning official launch of robotics foundation model in first half of 2026; corporate flip to relocate headquarters from South Korea to United States announced in February 2026
Market positioning
Physical AI challenger focused on industrial robotics intelligence; 'robotics brain' specialist enabling human-level dexterity and cognition for machines
Geographic focus
North America (Tesla, Figure AI, 1X headquarters), global humanoid robot market with differentiation through Korean and Japanese manufacturing ecosystem access
Patents and IP
No publicly disclosed patents or trademarks filed by RLWRLD as of current research. The company may be prioritizing trade secrets over public IP filings, which is common in early-stage AI/robotics startups.
About Jung-hee Ryu
Founder & CEO of RLWRLD (2024-present). Previously founded and led FuturePlay (2012-2024), a prominent Korean VC accelerator. Founded Olaworks (2006), the first Korean startup acquired by Intel (2012). KAIST PhD in Electrical Engineering and Computer Science, former Associate Professor.
Official website: https://www.rlwrld.ai/