AI Robot Control Startup Robai Secures Seed Funding from Korea Investment Accelerator

Seoul, July 28, 2026 – Robai, an AI robot control startup, announced today that it has secured seed investment from Korea Investment Accelerator. The company did not disclose the investment amount.

Robai creates not the robot’s “brain” but its “motor nerves,” enabling robots to autonomously learn their own body structure and movements. The company is developing a general-purpose robot control platform called “Robot Academy” based on “AI Self-Exploration” technology, which allows robots to independently explore and learn their joint structures, sensor configurations, movement ranges, and physical constraints. The goal is to solve the limitations of conventional development approaches that required redesigning control systems and rebuilding training data whenever robot types or hardware changed.

AI Self-Exploration consists of three models. The Robot Action Recognition (RAR) model automatically recognizes what the robot is currently doing, the Motor Babbling AI model autonomously attempts movements meaningful for learning, and the Body Schema model learns the relationship between body structure and movements based on this data. Through the iterative process of these three models exchanging learning results with each other, robots come to understand their own bodies without human intervention.

This technology serves to translate high-level behavioral commands generated by Robot Foundation Models (RFM) into actual robot movements. While RFMs have been developing rapidly in recent years, the execution stage still requires individual design and repeated tuning of control systems according to each robot’s structure and operation method. Robai’s strategy is to replace this gap with AI-based general-purpose control software.

Robai was founded by a team with accumulated R&D experience in robot AI and control technology. With this investment, the company will advance its self-exploration technology and proceed with technology verification based on physical robots and expansion of R&D personnel. Starting with industrial robots, the company plans to expand its application scope to service robots, humanoids, and soft robots, establishing a standard for AI control technology in the execution layer that connects robots and high-level AI.

“For the robot industry to spread in earnest, the structure that requires developing control software from scratch every time new hardware emerges must change first,” said Won-young Jeong, CEO of Robai. “We will accelerate the expansion of the robot market by building a general-purpose AI control infrastructure that enables robots to understand their own body structure and movements and quickly adapt to various sites.”

“While RFMs are rapidly automating high-level decision-making for robots, the final section that implements those decisions on actual hardware is still tied to manual human work,” explained Dong-heon Yoo, head of the investment division at Korea Investment Accelerator. “Robai directly targets this gap, and we expect the company to occupy a meaningful position in the physical AI transition period, given that it is a rare founding team capable of handling every layer from hardware to AI within a single team.”