Mikelo Robotics Targets Surface Treatment with Physical AI

Recap

Source: 디일렉, report of Oct. 2, 2026

According to the report, Mikelo Robotics is deploying physical AI solutions specifically for surface treatment processes, including painting, sanding, and polishing. The company states it has secured 21 major customer projects in sectors such as automotive, power generation, shipbuilding, and defense. Key clients include Hyundai Rotem, SL Corporation, Hankook Tire, and Wonik QNC. The firm aims to generate over 3 billion KRW in revenue this year.

The core technology involves converting skilled workers' movements and process data into AI-learned patterns that robots can execute. This addresses the difficulty of automating tasks where product shapes or process variables (like paint flow or bubbles) are inconsistent, which traditional fixed-path robots cannot handle. Mikelo offers three main solutions: 'Motion' for replicating human work patterns, 'Path Generation' for creating 3D-based work paths, and 'VisCo' for real-time path correction using 3D vision. The software is designed to be agnostic to robot manufacturers, allowing integration with existing industrial or collaborative robots.

In terms of financial performance, the company reported revenue growth from 210 million KRW in 2023 to 1.26 billion KRW in 2025. Automotive accounts for 34% of this revenue. In August, Mikelo became a subsidiary of KOSDAQ-listed company Finger, which acquired a 79.48% stake for approximately 16 billion KRW. The company plans to shift from direct system integration to a SaaS model by 2028, targeting 5 billion KRW in revenue by 2027 and 12 billion KRW by 2028.

Context

Surface treatment, particularly painting and polishing, has historically been one of the hardest areas to automate due to the high variability in product geometry and the need for fine motor skills. Traditional automation relies on fixed coordinates, which fails when parts are slightly misaligned or when the process involves fluid dynamics like spray painting. Physical AI, which uses AI to perceive and adapt to the physical world in real-time, is emerging as a solution to bridge this gap. By training on human expert data, these systems can replicate the nuanced adjustments a human worker makes, such as changing spray distance or angle to avoid drips.

The acquisition by Finger, a company with cloud infrastructure capabilities, signals a strategic shift towards software-as-a-service in industrial robotics. This aligns with a broader trend where robotics companies are moving away from selling hardware-centric systems to selling intelligent software layers that can run on various hardware platforms. This model reduces the barrier to entry for manufacturers who already own robots but lack the intelligence to optimize them for complex tasks.

Robot's take

Mikelo’s approach of decoupling the AI software from specific robot hardware is a smart move for market penetration. It allows them to leverage existing installed bases of industrial robots, reducing the capital expenditure required for customers. However, the transition to a SaaS model is ambitious. Industrial environments are often isolated from the cloud for security and latency reasons, so the feasibility of a pure SaaS model in heavy manufacturing remains a question mark. The company will need to demonstrate that its AI can generalize well across different product types without extensive retraining for each new client. The rapid revenue growth is promising, but scaling from 21 projects to a global platform will require significant investment in data collection and edge computing infrastructure.