RLWRLD, a South Korean “physical AI” startup building robotics foundation models trained inside live industrial sites, has raised about $26 million in a Seed 2 funding round, bringing its total seed financing to roughly $41 million, the company said.
The Seed 2 round drew a mix of venture and corporate-backed investors, reflecting a growing investor push to back AI systems that can move beyond software demos into factories, warehouses and other operational settings where deployment hurdles are higher and data is harder to collect.
Headline Asia and Z Venture Capital joined as financial investors, RLWRLD said, alongside strategic investors that include CJ Logistics, Kakao Investment, Lotte Ventures, Hanwha Asset Management, Hyosung Ventures, Smilegate Investment and T Investment, among others.
Hashed Ventures, which led the company’s earlier Seed 1 round, also participated in Seed 2, RLWRLD said.
RLWRLD is using the round not only to fund product development but also to deepen an industry consortium of strategic investors and partners that can host proof-of-concept projects and deployments.
The company said multiple proof-of-concept and “robotics transformation” projects are underway with partners in South Korea and Japan, including work tied to logistics and distribution environments.
Unlike robotics AI efforts trained primarily in lab settings, RLWRLD’s approach emphasizes collecting real-world, multimodal data directly from operating industrial sites through partner access, which it says can translate into an advantage when models face variability and edge cases in production.
The company plans to officially unveil its robotics foundation model in the first half of 2026, it said.
The funding comes as interest rises in “physical AI”, AI systems that perceive and act in the real world, amid labour shortages in parts of Asia and corporate efforts to automate tasks in warehouses, manufacturing lines and last-mile logistics.
Recent investment has increasingly shifted toward data, tooling and deployment partnerships that can support robotics at scale, rather than prototypes that struggle to generalise outside controlled environments.
RLWRLD’s round underlines a key theme in robotics AI: access to real industrial environments can matter as much as model architecture.
Strategic investors such as logistics and retail-linked groups can provide deployment venues and proprietary operational data, which may help the company iterate faster and build moats that are difficult for lab-only competitors to replicate.