Proceedings · Session S-699 · filed November 4, 2026
AI & Emerging Tech in R&DSession paper
RLWRLD hires IBM veteran as COO to scale robot foundation model
RLWRLD names IBM Consulting Korea's former managing partner Min Jung Kim as COO, tasked with turning its RLDX-1 robot model and $41M seed backing into paid deployments.
By Priya Raman3 min read576 words
Summary
- RLWRLD named former IBM Consulting Korea managing partner Min Jung Kim as COO.
- The company has raised about $41 million in seed funding, including $26 million in February.
- RLWRLD's technical report claims RLDX-1 achieved 86.8% task success on the ALLEX humanoid versus about 40% for pi0.5 and GR00T N1.6.
- Commercialization pilots are underway with industrial partners in South Korea, Japan, Europe and the United States.
- The announcement disclosed no figures on paid deployments or revenue.
Physical AI developer RLWRLD has named Min Jung Kim, former managing partner of IBM Consulting Korea, as chief operating officer, charging her with converting a robotics foundation model that the company says beat two high-profile rivals on a standardized manipulation test into paid industrial deployments.
The appointment, announced by the South Korean–headquartered company, puts Kim in charge of global operations at a business that has raised roughly $41 million in seed funding to date. That total includes a $26 million Seed 2 round in February, with backing from CJ Logistics, Lotte Ventures, Kakao Investment and Hanwha — a roster that signals where the company expects its first customers to come from.
RLWRLD builds AI models for robots with five-finger hands performing precise manipulation in manufacturing, logistics and service settings. Its flagship model, RLDX-1, debuted at Dexterity Night, a San Francisco showcase where the company demonstrated grasping, pouring and tool use.
What do the benchmark numbers actually show?
The company's technical report claims RLDX-1 succeeded on 86.8% of tasks run on WIRobotics' ALLEX humanoid. In the same evaluation, Physical Intelligence's pi0.5 and NVIDIA's GR00T N1.6 reached roughly 40%.
Those figures come from RLWRLD's own report, on one hardware platform, and the announcement offered no third-party validation or task-count breakdown. The comparison still matters for R&D managers tracking foundation models for dexterous manipulation, because Physical Intelligence and NVIDIA are the reference points most industrial robotics teams currently evaluate against.
RLWRLD says RLDX-1 rests on what it calls a Multi-Stream Action Transformer, combining cognition, action, proprioception and physical-sensing streams. Its data strategy is field-driven: robots deployed at industrial sites in South Korea and Japan record how skilled workers perform tasks, and that recordings feed back into model training.
Who is the new COO?
Kim brings a consulting-and-conglomerate background rather than a robotics resume. At IBM Consulting she led Asia-Pacific corporate strategy after serving as managing partner for the Korean business. Earlier roles span McKinsey in Korea and Germany, executive positions at SK and Hanwha, and most recently the CEO seat at LF Food.
Her academic training is in chemistry: a bachelor's degree from POSTECH and a doctorate in polymer chemistry from the University of Stuttgart.
Her mandate covers strengthening the operating structure and supporting strategic relationships with customers and industrial partners as the company scales — in effect, the unglamorous work of turning pilot projects into contracts.
Where does commercialization stand?
Commercialization pilots are underway with industrial partners in South Korea, Japan, Europe and the United States, according to the company. The announcement included no figures on paid deployments or revenue, so the state of monetization remains unverifiable from public disclosure.
That gap defines Kim's job. In the announcement, she framed the challenge in organizational rather than technical terms.
"Physical AI has the potential to fundamentally transform how industries operate, but realizing that potential requires more than technological breakthroughs," Kim said. "It requires a deep understanding of industrial challenges and the organizational capabilities to translate innovation into scalable business solutions."
For R&D portfolio watchers, the open question is whether a self-reported 86.8% success rate on one humanoid platform and four regional pilot programs can survive contact with procurement departments. The February Seed 2 round gives RLWRLD runway; the next signal worth tracking will be a disclosed paid deployment or an independently evaluated benchmark result.
via rlwrld.ai (Original)
Filed under
- physical-ai
- robotics
- foundation-models
- industrial-ai
- commercialization
More from Priya Raman
Show full bio
Staff writer covering money, funding and market shifts around research and development at Hypothesis Wire.
249 articles
References
- Moonshot's Kimi K3.1 Pro Nears Frontier at Half Opus Pricing
- AI Spending Boom Shapes the 2026 Innovation Rankings
- Opentrons Banks $40M Series E to Push AI-Run Lab Automation
- Google and Synaptics Preview Edge AI Microcontroller Ahead of Oct. 20 Launch
- Anthropic and Novo Nordisk expand Claude work into drug discovery