“On the factory floor, experienced operators often diagnose and resolve equipment issues based on years of accumulated know-how. While this approach has proven effective, imagine if AI could perform those diagnoses and recommendations more accurately and faster. Supporting frontline workers with AI and fundamentally transforming factory operations is the vision behind our AI Native Factory.”
In a recent interview, Sung Min-seok, Vice President of onepredict, shared the company's vision for the future of industrial AI. Founded by Professor Yoon Byung-dong of Seoul National University, onepredict has built its reputation on AI-powered predictive maintenance technologies and is now expanding beyond equipment diagnostics to intelligent manufacturing operations.
From Equipment Diagnostics to Plant-Wide Manufacturing Intelligence
Before joining onepredict, Sung led technology planning at Doosan Enerbility and oversaw IoT platform development and business scaling at SOCAR and 42dot. Since joining onepredict in 2024, he has been responsible for transforming the company's core AI technologies into enterprise software solutions for industrial customers.
Under his leadership, onepredict's order volume grew from approximately KRW 700 million in 2024 to KRW 10 billion in 2025, with KRW 9 billion generated by pdx alone.
pdx is an AI-powered industrial asset management platform that enables manufacturers to manage multiple assets at the process and production-line level. Since its launch in 2025, the platform has rapidly expanded its customer base and become onepredict's flagship growth engine.
Unlike conventional monitoring systems that analyze equipment data individually, pdx integrates information from equipment, manufacturing processes, and factory operating systems to identify root causes and improve production efficiency. By demonstrating measurable improvements in product quality and manufacturing yield, the platform has seen rapid adoption across industrial sectors. One representative customer is Nongshim, which is currently conducting pilot testing to deploy the solution at its new manufacturing facility.
According to Sung, while a typical industrial AI deployment requires around 13 months, onepredict's extensive industrial data and pre-trained AI models enable customers to begin operating AI after only two to four weeks of on-site retraining.
Today, onepredict's solutions are deployed across a wide range of industries, including railway maintenance, defense, and data center HVAC systems. In 2025, the company also supplied its technology to Motiva Enterprises, the Saudi Aramco subsidiary and North America's largest refinery operator.
Sung explained that AI has fundamentally changed factory operations. Whereas identifying the root cause of equipment failures previously required specialists to spend weeks investigating on-site, AI can now generate analysis reports immediately and issue work orders directly to operators, significantly reducing downtime and maintenance costs.
Building Industrial Sovereign AI
onepredict also aims to realize the vision of Sovereign AI for industrial environments.
As global technology companies continue to dominate general-purpose AI, the company believes that relying on overseas cloud infrastructure for sensitive manufacturing and energy data presents significant security risks for national critical industries.
According to Sung, Sovereign AI is not simply about developing domestic AI models—it is about maintaining control over industrial data and AI operations throughout the manufacturing lifecycle. onepredict's technology collects equipment data, performs AI analysis, and supports operational decision-making directly at manufacturing sites, providing a practical model for industrial Sovereign AI.
Unlike conventional AI solutions that depend on externally hosted data, onepredict's domestically developed platform generates AI-ready data directly from industrial equipment. By continuously producing and processing data within the manufacturing environment, the platform helps prevent sensitive factory data from leaving the site.
Sung compared today's industrial AI to Advanced Driver Assistance Systems (ADAS) in the automotive industry, where AI assists operators by providing alerts and recommendations. The company's long-term vision is to evolve toward fully autonomous manufacturing, where AI independently makes operational decisions and resolves issues without human intervention—ultimately reaching the equivalent of Level 5 autonomous operation.
Looking ahead, onepredict plans to accelerate its global expansion beyond Korea. In addition to North America and Vietnam, the company is actively evaluating opportunities in Japan, where industrial AI adoption is gaining momentum. By first supporting Korean manufacturers with overseas production facilities, onepredict aims to rapidly build global references and expand its international presence.
...
...
...
...
In a recent interview, Sung Min-seok, Vice President of onepredict, shared the company's vision for the future of industrial AI. Founded by Professor Yoon Byung-dong of Seoul National University, onepredict has built its reputation on AI-powered predictive maintenance technologies and is now expanding beyond equipment diagnostics to intelligent manufacturing operations.
From Equipment Diagnostics to Plant-Wide Manufacturing Intelligence
Before joining onepredict, Sung led technology planning at Doosan Enerbility and oversaw IoT platform development and business scaling at SOCAR and 42dot. Since joining onepredict in 2024, he has been responsible for transforming the company's core AI technologies into enterprise software solutions for industrial customers.
Under his leadership, onepredict's order volume grew from approximately KRW 700 million in 2024 to KRW 10 billion in 2025, with KRW 9 billion generated by pdx alone.
pdx is an AI-powered industrial asset management platform that enables manufacturers to manage multiple assets at the process and production-line level. Since its launch in 2025, the platform has rapidly expanded its customer base and become onepredict's flagship growth engine.
Unlike conventional monitoring systems that analyze equipment data individually, pdx integrates information from equipment, manufacturing processes, and factory operating systems to identify root causes and improve production efficiency. By demonstrating measurable improvements in product quality and manufacturing yield, the platform has seen rapid adoption across industrial sectors. One representative customer is Nongshim, which is currently conducting pilot testing to deploy the solution at its new manufacturing facility.
According to Sung, while a typical industrial AI deployment requires around 13 months, onepredict's extensive industrial data and pre-trained AI models enable customers to begin operating AI after only two to four weeks of on-site retraining.
Today, onepredict's solutions are deployed across a wide range of industries, including railway maintenance, defense, and data center HVAC systems. In 2025, the company also supplied its technology to Motiva Enterprises, the Saudi Aramco subsidiary and North America's largest refinery operator.
Sung explained that AI has fundamentally changed factory operations. Whereas identifying the root cause of equipment failures previously required specialists to spend weeks investigating on-site, AI can now generate analysis reports immediately and issue work orders directly to operators, significantly reducing downtime and maintenance costs.
Building Industrial Sovereign AI
onepredict also aims to realize the vision of Sovereign AI for industrial environments.
As global technology companies continue to dominate general-purpose AI, the company believes that relying on overseas cloud infrastructure for sensitive manufacturing and energy data presents significant security risks for national critical industries.
According to Sung, Sovereign AI is not simply about developing domestic AI models—it is about maintaining control over industrial data and AI operations throughout the manufacturing lifecycle. onepredict's technology collects equipment data, performs AI analysis, and supports operational decision-making directly at manufacturing sites, providing a practical model for industrial Sovereign AI.
Unlike conventional AI solutions that depend on externally hosted data, onepredict's domestically developed platform generates AI-ready data directly from industrial equipment. By continuously producing and processing data within the manufacturing environment, the platform helps prevent sensitive factory data from leaving the site.
Sung compared today's industrial AI to Advanced Driver Assistance Systems (ADAS) in the automotive industry, where AI assists operators by providing alerts and recommendations. The company's long-term vision is to evolve toward fully autonomous manufacturing, where AI independently makes operational decisions and resolves issues without human intervention—ultimately reaching the equivalent of Level 5 autonomous operation.
Looking ahead, onepredict plans to accelerate its global expansion beyond Korea. In addition to North America and Vietnam, the company is actively evaluating opportunities in Japan, where industrial AI adoption is gaining momentum. By first supporting Korean manufacturers with overseas production facilities, onepredict aims to rapidly build global references and expand its international presence.