Korean industrial giants back Config in $27m seed round as it targets the data layer for robotic AI
The oversubscribed funding round values Config at over $200 million as the firm scales its collection of human motion data for robotic foundation models.
South Korea's leading manufacturing conglomerates have collectively invested in Config, a Seoul-based startup positioning itself as the essential data infrastructure for the next generation of robotics. In a strategic move that mirrors the semiconductor industry's reliance on foundries, Samsung Venture Investment led an oversubscribed $27 million seed round that values Config at more than $200 million.
The investment group includes strategic participation from Hyundai Motor's ZER01NE Ventures, LG Tech Ventures, and SKT America, alongside financial backers such as Mirae Asset Ventures and Korea Development Bank. This capital injection marks a significant shift where major hardware manufacturers are backing the data layer required for robotic foundation models rather than solely building proprietary robots themselves.
Founded in January 2025 by CEO Minjoon Seo, a former researcher at Meta and chief scientist at Twelve Labs, the company has already established a substantial operational footprint. Config currently employs nearly 300 staff members across its offices in Seoul and Hanoi, Vietnam, where it has accumulated over 100,000 hours of human motion data. This dataset is described as more than 30 times the size of AgiBot World, the largest comparable open-source dataset in the sector.
Unlike software-only artificial intelligence, which relies on vast amounts of text that are relatively easy to obtain, teaching robots to move presents a unique physical challenge. Every piece of training data must be physically collected, requiring robots, facilities, and human operators. Config addresses this by transforming raw human motion data before it is used for training, a process the company compares to language translation to ensure the information is optimally suited for robotic movement.
The startup is already generating revenue from large manufacturers, system integrators, and sectors including agriculture and defence. The new funding will be directed toward three key priorities: scaling data operations to one million hours by 2027, achieving $10 million in annual recurring revenue by the end of that year, and launching a cloud-based Robot-as-a-Service product that allows companies to run the foundation model without onboard hardware.
