OnePredict outlines an AI-native factory vision built on data unification and multi-agent coordination, not humanoid robots.
The AMW Read
Korean industrial-AI vendor repositions from point predictive-maintenance tools to an orchestration-OS narrative for factory-wide physical AI, without a funding event, new product launch, or named-player disruption.
OnePredict outlines an AI-native factory vision built on data unification and multi-agent coordination, not humanoid robots.
At a webinar held August 31, OnePredict COO Sung Min-seok argued that physical AI in manufacturing is not a new technology but a convergence of existing ones — sensors, predictive-maintenance models, and pre-programmed robots — now linked through foundation models and Vision-Language-Action (VLA) models that connect perception to action. He said the shift is from rigid, rule-based automation to adaptive systems that interpret context and choose actions, and introduced the company's data platform Cyclone, which unifies operational- and information-technology data across PLCs, DCS, and MES systems, alongside its AI application platform PDX, which applies that data to maintenance and quality problems.
The framing matters because it pushes physical AI's definition well past humanoids: OnePredict counts AMRs, drones, CNC machines, compressors, and turbines, plus any action that adjusts a valve, pump speed, or setpoint, as physical AI in scope. Sung proposed a coordination layer he called factory intelligence, run by an AI-native operating system that arbitrates when a maintenance agent, a production agent, and a quality agent issue conflicting recommendations — for example, a 12-hour repair window against a 6-hour production extension. Human oversight persists through graduated human-in-the-loop, human-on-the-loop, and human-above-the-loop control levels rather than full removal of human judgment.
For builders and investors, the pitch signals that industrial AI vendors are repositioning from point predictive-maintenance products toward owning the orchestration layer that will matter once agentic and VLA-capable systems reach factory floors. The monetizable step today is still OT/IT data integration and maintenance-and-quality use cases through Cyclone and PDX — autonomous factory-wide execution remains a stated destination, not a shipped capability.
#PhysicalAI #IndustrialAI #ManufacturingAI #AIAgents #OnePredict #FactoryAutomation