Whalesbook
Config's Seed Round Fuels Robot Data Infrastructure
Config's substantial seed funding round is a strong signal of support for its mission: building the foundational data layer for the next generation of artificial intelligence, known as embodied AI. With robots increasingly used for physical tasks, the challenges and opportunities in collecting data for these systems are becoming critical, especially in Asia's manufacturing hubs.
Asia's Industrial Strength Powers Robotics AI
Config's backing by leading South Korean manufacturers like Samsung, Hyundai, and LG highlights Asia's deep manufacturing expertise. Unlike AI for software, which relies on abundant digital data, training robots demands expensive, real-world physical data. This need makes the region's industrial infrastructure and supply chains ideal for advancing physical AI. Config's focus on the data layer for robotic foundation models directly tackles this challenge, aiming to boost robot capabilities without building the robots themselves.
Industry Giants Invest in Robotics Data Backbone
The $27 million seed round, led by Samsung Venture Investment and including ZER01NE Ventures (Hyundai Motor's arm), LG Tech Ventures, and SKT America, shows a unified industrial strategy. These major companies are investing not just in a startup, but in the infrastructure needed for their robotics projects and the wider physical AI field. South Korea's government also backs this, planning major investments in industrial AI, smart factories, and robotics to boost national competitiveness amid demographic shifts. This broad support confirms Config's role in supplying the essential elements for the fast-growing physical AI sector.
Config Aims to Be the 'TSMC of Robot Data'
Config sees itself as the "TSMC of robot data," similar to how the Taiwanese chipmaker manufactures for many companies without direct competition. By concentrating solely on the data layer for robotic foundation models, Config aims to be a vital partner for robot developers and businesses in areas like agriculture and defense. CEO Minjoon Seo notes that unlike text-based AI with vast data, robot training needs physical data from robots, environments, and people. This data must then be processed for robot learning, a key technical advantage for Config.
Robotics AI Data Market Sees Growing Competition
The market for robotics AI infrastructure is growing quickly, with substantial investment in developers of core AI models. Competitors such as Skild AI have reached valuations over $14 billion after large funding rounds, including a $1.4 billion Series C in early 2026. Physical Intelligence, which develops general AI models for robots, is reportedly discussing a $1 billion funding round at an $11 billion valuation as of March 2026. Other companies focused on embodied robotics intelligence have also attracted significant capital. While Config is in an earlier seed stage with a $200 million valuation, its focus on the data layer and strong industrial backing could secure a key market position. This sector saw about $5.3 billion invested in physical AI in April 2026 alone.
Bridging the Robotics Data Gap
Config's main offering tackles the vast difference between data available for software AI and what robots need. Large language models (LLMs) train on tens of billions of hours of text, but high-quality data for embodied AI is estimated at only 500,000 hours globally. Gathering robot data is difficult and costly, requiring physical robots, specific environments, and staff. Current industry data collection often yields only 25-37%. Config's approach involves collecting and processing over 100,000 hours of human motion data to ease this scarcity and help more advanced robots be used. The company plans to expand data operations to one million hours and introduce a Robot-as-a-Service cloud product.
Challenges Ahead: Execution and Competition
Despite its strategic placement, Config faces significant challenges. The company's goals to scale data operations to one million hours and reach $10 million in Annual Recurring Revenue by the end of 2026 demand perfect execution in varied locations like Vietnam and Seoul. The main hurdles for robotics AI data—its scarcity, expense, and need for complex formats (like detailed hand data)—remain substantial. Competition is intense, with well-funded rivals and tech giants like NVIDIA creating AI models and simulations to speed up development. Any error in data quality, collection efficiency, or platform growth could allow competitors with more resources or established data systems to gain an advantage. Additionally, reliance on physical hardware adds inherent complexity and potential delays.
Robotics Market Growth Fuels Config's Opportunity
The global robotics market is forecast to hit $38 billion in 2026, showing its fastest growth in ten years. Config's recent seed funding positions it to be a key provider of the data infrastructure needed for this expansion. The company's plan to make high-quality robotics data more accessible via its platform and Robot-as-a-Service offering fits with the growing use of industrial AI, enabling Config to benefit from the increasing demand for sophisticated robotic systems.