Korea’s largest manufacturers support Config, which is TSMC’s robot data hub

Asia’s push toward physical AI is fueled by the same manufacturing prowess that has made the region a global industrial powerhouse. Across South Korea, Japan, China and Taiwan, manufacturing remains a key pillar of economic growth. Unlike more service- or software-oriented economies, these countries have long relied on large-scale production, export-based industries, and highly improved supply chains. This structural foundation is now shaping how AI is adopted and where investments flow.

Which makes it especially significant that Config, a Seoul and San Jose-based startup building the data layer for robotic foundation models (RFMs), has secured support from the venture arms of South Korea’s largest manufacturers.

Samsung Venture Investment led its oversubscribed $27 million seed round at a valuation of more than $200 million, bringing Config’s total to $35 million. ZER01NE Ventures, Hyundai Motor’s venture arm, LG Tech Ventures, and SKT America, the venture capital unit of the South Korean telecom giant, also joined as strategic investors, along with angel investor Pieter Abbeel (covariant AI co-founder and UC Berkeley professor) and financial backers including Mirae Asset Ventures, Korea Development Bank, GS Futures, Kakao Ventures, and Z Ventures.

Config was founded in January 2025 by CEO Minjoon Seo, a former Meta researcher and chief scientist at Twelve Labs, along with three co-founders with backgrounds at Waymo, Google, and Naver. Instead of building the robots themselves, the team is focusing on a simpler goal: providing the data the robots need to learn and work. They believe that better data will be key to making robots more useful.

Training large language models is expensive, due to the computational power required to process them, but the raw materials, huge amounts of text, are easy to obtain online. Teaching robots how to move is a completely different challenge, Siu said in an exclusive interview with TechCrunch. Every piece of training data must be physically collected, such as your need for the robot, the facility running it, and the people operating it. This makes developing AI for bots more expensive than developing just chatbots, according to Seo. As companies seek to develop more capable robots, the cost of collecting and sorting data could rise quickly.

Config wants to be the company that makes everyone else’s automated AI possible. The startup compares its role to TSMC, the Taiwanese chipmaker that makes for Apple, Nvidia and AMD without competing with any of them. Config aims to play a similar role in robotics by providing data. This approach is gaining momentum as major manufacturers increasingly seek to build their own AI-powered robots rather than rely entirely on third-party vendors. This is the composition of the market he is betting on.

Config is already generating revenue, said Jack Pang, COO and co-founder of Config. The startup’s current clients include major manufacturers, systems integrators, and companies in the agriculture and defense sectors, Pang told TechCrunch. Their peers in this field include Physical Intelligence, Artificial General Intelligence, and Skilled AI.

The composition records humans performing physical tasks in controlled studio environments and in the field. The startup operates in Seoul and Hanoi, where a workforce of about 300 people produces the data. To date, it has collected more than 100,000 hours of human movement data, more than 30 times the size of AgiBot World, the largest comparable open source dataset at about 3,000 hours.

Most robotics teams train AI models on human movement data and then adapt those models to fit the robot. Genesis takes a different approach, Seo said. The company is focusing on transforming data before training begins so that it is more relevant to the way robots move and interact with the world. Seo compared the process to language translation. Training a model on one type of data and expecting it to work smoothly in another environment is an attempt to teach Korean using only English-language materials, Seo said.

“The data needs to be transformed, not the model. This transformation technology is Config’s core technical differentiator,” Seo said.

The funding will go toward three priorities: scaling its data operations in Vietnam and Seoul toward 1 million hours of collected data, growing its enterprise platform business to $10 million in ARR by the end of 2027, and launching a cloud-based Robot-as-a-Service product that allows companies to run Config’s foundation model without the need for embedded hardware.