Over 60 Clients Including Power Generation, Petrochemicals, and Heavy Industry
Launch of AI-Native Factory Solutions Such as pdx
Targeting IPO in the First Half of 2027

"If the past 10 years were an era of technological proof, the future will be an era of 'demonstration and diffusion' where industrial AI takes root," he said.


Minseok Sung, Vice President of Onepredict, described this inflection point in the industrial AI era during an interview with The Asia Business Daily on March 10, 2026. He said, "As the industrial AI market is just opening up, technology is accelerating, and adoption by companies will become standardized." Onepredict provides a platform that uses AI to prevent equipment failures in factories and other facilities, thereby increasing productivity.


[AI Revolution](190) Onepredict: "The Era of Industrial AI 'Proof' Is Over... Now Comes 10 Years of 'Diffusion'" View original image

Onepredict was founded in 2016 by CEO Byungdong Yoon, a former professor in the Department of Mechanical Engineering at Seoul National University, and developed a predictive maintenance solution that forecasts industrial equipment failures using AI. After working at Doosan Enerbility, Socar, and 42dot, Vice President Sung joined Onepredict and now leads the business as Chief Product Officer (CPO), overseeing the entire product lifecycle including planning, marketing, and project management.


Onepredict is a specialized industrial AI company that monitors industrial equipment and manages industrial data. The company has supplied solutions that reduce losses caused by factory shutdowns. For example, the "GuardiOne Turbo" applied to Company A's centrifugal compressor detected abnormalities and identified the cause of abnormal vibrations, preventing equipment downtime losses worth approximately 1 billion won.


Onepredict's core competitive edge lies in its industry-specialized AI foundation model. A foundation model is designed to learn industry-specific data, such as those from the energy and manufacturing sectors. A significant portion of its cumulative investment of 49 billion won was used to secure these models. CEO Yoon was able to build foundation models using the know-how he gained from researching both AI and mechanical engineering, as well as large-scale data. Vice President Sung explained, "Industrial AI requires a high level of mechanical engineering domain knowledge, and we have a foundation model that works in practice," adding, "It only takes about two to four weeks to activate a customer's data."


Recently, Onepredict introduced an AI-native factory solution that can be expanded beyond single industrial equipment to cover entire processes. The flagship products are "Cyclone" and "pdx." Vice President Sung said, "pdx can expand its scope to support connected equipment components—such as processes, lines, and even large-scale systems," and added, "Shortly after its launch last year, we secured orders worth about 5 billion won, and this year we expect at least double that amount."


The company has also been selected for the Ministry of Trade, Industry and Energy's AI Factory Leading Project and plans to implement its solution in food and beverage production processes. Vice President Sung explained, "When a factory is operated mainly by workers, equipment breakdowns can cause a complete production halt and require all materials to be discarded," and continued, "By applying our solution from raw material supply to packaging, we can address problems that arise during the production process."


Onepredict’s clientele includes around 60 companies such as POSCO, GS Caltex, GS Power, Korea Midland Power, as well as power plants, petrochemical, and heavy industry firms. Starting this year, the company also plans to expand into data centers, environmental safety, and other fields.



Onepredict is focusing on building its business performance with the goal of an initial public offering (IPO) in the first half of 2027. The company aims to grow into the leading industrial AI company in Korea. Vice President Sung said, "Until now, industrial AI has been defined in different ways, but going forward it will evolve and be categorized by company type," adding, "We have laid the groundwork to represent Korean industrial AI by overcoming various constraints, and will prepare for listing based on the business results generated by our technology."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing