Samsung, Hyundai, and LG back Config, a startup supplying training data for robots

South Korea's biggest industrial companies are investing heavily in a startup that supplies training data to robots. Samsung Venture Investment led Config's $27 million seed round at a valuation exceeding $200 million, bringing the company's total raised to $35 million. Hyundai Motor's venture arm ZER01NE Ventures, LG Tech Ventures, and SKT America also invested as strategic backers, alongside venture firms Mirae Asset Ventures, Korea Development Bank, GS Futures, Kakao Ventures, and Z Ventures.

Config, based in Seoul and San Jose, was founded in January 2025 by CEO Minjoon Seo, a former Meta researcher and ex-chief scientist at Twelve Labs. Three other co-founders came from Waymo, Google, and Naver. The company does not build robots itself. Instead, it collects, processes, and sells the human motion data that robot AI systems need to learn how to move and perform tasks.

Teaching robots to move presents a challenge unlike training software-only AI systems. Text data for language models is abundant and free across the internet, but robot training data must be physically collected. According to Seo, this requires actual robots, facilities to run them, and people to operate them. That makes robotics AI significantly more expensive to develop than chatbot software. As manufacturers demand more capable robots, the cost of gathering and labeling motion data grows quickly.

Config positions itself as critical infrastructure for the robotics industry, similar to how Taiwan's TSMC manufactures chips for Apple, Nvidia, and AMD without competing with them. "Config wants to be the company that makes everyone else's robot AI possible," the company explains. Large manufacturers increasingly prefer to build their own proprietary robot AI rather than depend entirely on outside vendors—a trend Config is betting on.

The startup already generates revenue from customers including large manufacturers, system integrators, and companies in agriculture and defense sectors, according to COO and co-founder Jack Bang. Config operates data production facilities in Seoul and Hanoi with a workforce of nearly 300 people. The company has accumulated over 100,000 hours of human motion data—more than 30 times the size of AgiBot World, the largest comparable open-source dataset at roughly 3,000 hours.

Config's approach to data preparation differs from standard practice in robotics. Most teams train AI models on human motion data and then adapt those models for robots, but Seo said Config transforms the data before training so it suits how robots actually move and interact with the world. "The data must be converted, not the model. This conversion technology is Config's core technical differentiator," Seo said. He compared the approach to language translation: training a model on one type of data and expecting it to work in another setting is like trying to teach Korean using only English-language materials.

Config will use the new funding for three main goals. First, it aims to scale its data operations in Vietnam and Seoul to reach one million hours of collected motion data. Second, it plans to grow its enterprise platform business to $10 million in annual recurring revenue by the end of 2027. Third, it will launch a cloud-based Robot-as-a-Service product allowing companies to run Config's foundation model without requiring hardware on their own premises.