East Asian industrial giants are quietly funding the backbone of the next AI revolution. Korea’s top conglomerates—Samsung, Hyundai, LG, and SK—have collectively poured millions into Config, a Seoul- and San Jose-based startup building the foundational data layer for robotic foundation models. Led by Samsung Venture Investment in a $27 million oversubscribed seed round, the capital infusion also includes Hyundai Motor’s ZER01NE Ventures, LG Tech Ventures, and SKT America. This strategic backing positions the company as the TSMC of robot data, creating a specialized supply chain that will accelerate physical AI adoption across industries.

Why Manufacturing Giants Are Betting on Physical AI

The transition from software-only AI to tangible robotics requires a completely different approach to scaling. Unlike text-based large language models that rely on relatively cheap digital datasets, training physical robots demands massive amounts of real-world motion data. Every movement must be physically captured, labeled, and contextualized under both controlled and field conditions.

As CEO Minjoon Seo noted in an exclusive interview, "Every piece of training data has to be physically collected... That makes robotics AI more costly than software-only models." Config bridges this expensive bottleneck by transforming raw human motion into robot-optimized formats before the AI training phase even begins. This technical differentiator often operates like a language translator for robotics, ensuring seamless integration across hardware manufacturers.

Config’s Core Technology and Strategic Advantages

The startup’s technical approach revolves around three primary pillars designed to streamline the path from motion capture to deployment:

  • Data Conversion Specialization: Rather than simply aggregating raw datasets, the company refines human motion into formats that align with specific robot kinematics.
  • Vertical Integration: By managing both data production through a dedicated workforce in Seoul and Hanoi and the underlying AI infrastructure, the firm maintains strict quality control while scaling efficiently.
  • Enterprise Platform Growth: The new capital will directly support a push toward $10 million in annual recurring revenue by 2026.

This funding directly expands the company's Robot-as-a-Service platform, which lets manufacturers run advanced models without investing in expensive onboard hardware. The approach offers a compelling proposition for industrial clients seeking rapid prototyping cycles without heavy upfront capital expenditure.

Becoming the TSMC of robot data in a Crowded Market

Config’s market positioning deliberately mirrors Taiwan’s semiconductor foundry model, but applied to physical automation. By operating as a neutral supplier, the startup enables diverse companies to build proprietary AI systems without relying on a single vendor’s ecosystem. This neutrality stands in stark contrast to competitors like Physical Intelligence or Skild AI, which typically focus on narrow robot types or specific industrial verticals.

Targeting a universal data backbone allows Config to attract enterprise partners wary of vendor lock-in while simultaneously pushing for standardization across the broader robotics ecosystem. Acting as the de facto TSMC of robot data requires maintaining strict neutrality across competing hardware manufacturers. The funding also directly supports the expansion of its Robot-as-a-Service platform, which lets manufacturers run advanced models without investing in expensive onboard hardware.

The Road Ahead for Physical AI

With the fresh capital directed toward scaling operations to one million hours of collected data, Config is betting heavily on the viability of standardized datasets. As South Korea’s industrial leaders accelerate their digital transformation, their financial commitment signals a critical pivot where traditional manufacturing expertise meets frontier robotics research. The convergence of massive capital, advanced engineering, and industrial scale suggests that robot AI will rapidly transition from experimental prototypes to production-grade solutions.

Whether this model achieves long-term sustainability will depend on delivering measurable ROI for enterprise clients while navigating the complex ethical frameworks surrounding large-scale physical data collection. The company must also ensure its data pipelines remain secure and compliant across global jurisdictions. Ultimately, the support from these heavyweight investors underscores a shared belief that East Asian manufacturing dominance is rapidly expanding from physical assembly lines into the very code powering tomorrow’s intelligent machines.