[For more diverse corporate information on the startups mentioned in this article, please visit Unicorn Factory's big data platform 'Data Lab'.]

Major domestic conglomerates, including Samsung, Hyundai Motor, LG, and SK, along with other large corporations, have collectively poured capital into a new startup developing the Robot Foundation Model (RFM). Config Intelligence announced on the 17th that it secured 40 billion won at the seed stage by attracting interest from major corporate venture capital (CVC) arms of these conglomerates, leveraging its AI model technology that serves as a kind of "robot brain."
This round was led by Samsung Venture Investment, with participation from Hyundai Motor's Zero One Ventures, LG Technology Ventures, SK Telecom America, Kakao Ventures, GS Futures, and Z Venture Capital (a joint venture between Naver and SoftBank). Also listed among the investors were Mirae Asset Venture Investment, Korea Development Bank, and UC Berkeley Professor Peter Abbeel, co-founder of Covariant. It is reported that most of these investors also participated in the pre-seed stage. The cumulative investment amount, including the pre-seed round, stands at 50 billion won.
The primary reason cited by investors for betting on Config Intelligence is RFM's technical approach. RFM is a general-purpose intelligent control model that understands and learns from diverse physical environments and tasks based on large-scale data and general AI, enabling it to be applied across multiple robots.
According to investors, the company is developing RFM by combining the "world model" approach, which generates video to predict future scenarios, with the "VLA (Vision-Language-Action)" approach, which translates cognition into action. This hybrid methodology has been highly rated for its technical maturity, as it can deliver stable performance across a wider range of environments compared to single-method approaches.
Data competitiveness was also a core rationale behind the investment decision. Config Intelligence directly operates a factory subsidiary in Vietnam with a workforce of over 300 people, securing more than 100,000 hours of human motion data. This figure is more than 33 times larger than the robot learning dataset (3,000 hours) released as open source by China's Agibot.
Shin Jeong-ho, Senior Reviewer at Kakao Ventures, stated, "What sets them apart is not merely having a large volume of data, but the rare technical capability to refine that raw material into a form that actual robots can learn from." He added, "A team that has internalized this pipeline is among the few globally recognized for such expertise."
He further remarked, "Config started by solving the critical bottleneck of high-quality data pipelines in the era of physical AI and has built the technical capability to expand into a robot-specific AI model (RFM) capable of delivering high performance across diverse environments. With the strategic participation of global manufacturers like Samsung and Hyundai Motor, we expect them to generate meaningful synergy in actual field operations and grow as Korea's leading AI robotics company."
The founding team's credentials also contributed significantly to investor confidence. Seo Min-jun (CEO) is an AI expert who has worked across major tech giants such as Naver, Meta, and Twelve Labs, as well as startups, and currently serves as a professor at the Graduate School of Artificial Intelligence at KAIST. Chief Technology Officer (CTO) Son Hyung-mok holds a Ph.D. in physics from Harvard University and previously improved and validated cognitive system performance, including vision sensors for autonomous driving systems, at Waymo. Chief Science Officer (CSO) Lee Gi-min has conducted pioneering research in robotics and AI, having worked at the UC Berkeley Artificial Intelligence Research Institute (BAIR) and Google before serving as a part-time professor at KAIST.

Seo (CEO) plans to utilize the funds raised from this investment round to advance the RFM model. The largest portion of costs will be allocated to R&D (research and development) personnel expenses. Funds will also be used for GPU training costs, robot procurement, and other related expenditures. A certain portion of the budget is also planned for data collection costs. Seo (CEO) stated, "Since video data is close to raw material and difficult to utilize without refinement, we plan to concentrate investment funds on advancing technology in the data refinement stage and conducting verification tasks to ensure stable operation of the actual model."
Config Intelligence does not manufacture robot hardware directly. Instead, it aims to accelerate the commercialization of cloud-based Robot-as-a-Service (RaaS) designed for use with robots employed by physical AI companies or factories operating facilities. TechCrunch has described Config Intelligence as the "TSMC of the robotics world," meaning a company that supplies data infrastructure enabling other firms to deploy robot AI, much like a foundry responsible for semiconductor manufacturing.
[MoneyToday startup media platform Unicorn Factory]