Korea’s biggest manufacturers back Config, the TSMC of robot data

Source: TechCrunch

Background

Asia’s push into physical AI is being fueled by the same manufacturing prowess that made the region a global industrial powerhouse. Across South Korea, Japan, China and Taiwan, manufacturing remains a central pillar of economic growth. Unlike economies more heavily weighted toward services or software, these countries have long relied on large‑scale production, export‑driven industries, and highly optimized supply chains. That structural foundation is now shaping how artificial intelligence is adopted and where investment flows.

Funding Round

Config, a Seoul‑ and San Jose‑based startup building the data layer for robotic foundation models (RFMs), secured backing from the venture arms of South Korea’s biggest manufacturers.

• Lead investor: Samsung Venture Investment
• Round size: $27 million seed round (oversubscribed)
• Valuation: > $200 million
• Total capital raised: $35 million

Other strategic investors include

• Hyundai Motor’s ZER01NE Ventures
• LG Tech Ventures
• SKT America (telco giant’s VC unit)
• Angel investor Pieter Abbeel (co‑founder of Covariant AI, UC Berkeley professor)
• Mirae Asset Ventures, Korea Development Bank, GS Futures, Kakao Ventures, Z Ventures

Company Overview

Founded in January 2025 by CEO Minjoon Seo (former Meta researcher and chief scientist at Twelve Labs) and three co‑founders with backgrounds at Waymo, Google, and Naver, Config focuses on providing the data robots need to learn and operate rather than building robots themselves.

Key points from an exclusive interview with TechCrunch:

• Training large language models is expensive mainly due to compute, but the raw text data is abundant.
• Teaching robots to move requires physically collected data—robots, facilities, and operators—making robotics AI costlier than software‑only chatbots.
• As robots become more capable, data collection and labeling costs can rise quickly.

Business Model & Market Position

Config positions itself as the “TSMC of robot data,” supplying high‑quality training data to robot AI developers without competing with them. The startup’s customers already include large manufacturers, system integrators, and firms in agriculture and defense.

Peers in the space

• Physical Intelligence
• Generalist AI
• Skild AI

Data Collection & Scale

Image Credit: Kate Park

• Operates out of Seoul and Hanoi with a workforce of nearly 300 handling data production.
• Accumulated over 100,000 hours of human motion data—more than 30 × the size of the open‑source AgiBot World dataset (~3,000 hours).

Approach

Most robotics teams train AI models on raw human motion data and then adapt them for robots. Config instead transforms the data before training, making it better suited to robot kinematics and interaction. Seo likens this to language translation: “Training a model on one type of data and expecting it to work seamlessly in another setting is like trying to teach Korean using only English‑language materials.”

“The data must be converted, not the model. This conversion technology is Config’s core technical differentiator.” — Minjoon Seo

Future Plans

The new funding will be allocated to three priorities:

1. Scale data operations in Vietnam and Seoul toward one million hours of collected data.
2. Grow the enterprise platform to reach $10 million ARR by the end of 2027.
3. Launch a cloud‑based Robot‑as‑a‑Service (RaaS) product, enabling companies to run Config’s foundation model without onboard hardware.