Why Data Format Is the Problem Now
On most SME shop floors, every machine and process speaks its own data dialect. The same temperature reading carries a different name, unit, and sampling interval from one machine to the next, so each new AI solution needs its own custom interface. The Manufacturing Data Standardization Program, for which the Ministry of SMEs and Startups and the AI Innovation Task Force at the Korea Technology and Information Promotion Agency for SMEs (TIPA) held a kickoff meeting on July 1, 2026, targets exactly this problem. Its aim is to unify data around international standards so that AI can be applied on site without separate development work.
Earlier discussions around OPC UA were about communication: how data gets exchanged. This program standardizes the asset information model itself: what data is captured, and in what structure.
The Program in Numbers
Note that 500 is a year-end target. Whether it is met, and how widely the models are released, should be confirmed against the program's official results.
What Is an AAS? The Executive-Level Version
The Asset Administration Shell (AAS) is a standard model for digitally representing and managing information about physical assets such as equipment, and it is defined in the international standard IEC 63278-1:2023. The key concept is the submodel. Nameplate data (manufacturer, model, serial number), technical data (specifications), and operational data (temperature, vibration, running status) each go into a predefined structure. A reference model is a ready-made template of that structure for a specific type of equipment or process.
AAS does not replace existing technologies. OPC UA is the communication layer that carries data, MES is the business system that manages production, and a digital twin is an application that uses the data. AAS sits between them and defines what the data means and how it is organized.
What Changes for Your Plant
Data Standards Checklist Before Ordering a Smart Factory
Five Questions to Ask Your Supplier
Common Misconceptions and Cautions
Connecting It to Government Support Programs
If you are preparing for 2027 smart factory or manufacturing AI programs, build a concrete data standardization plan into your business plan. Linking target equipment, data items, the standards and reference models to be applied, and post-build uses such as quality prediction or predictive maintenance into a single storyline makes feasibility easier to demonstrate. Data integration across machines and systems matters more as you move through level verification and into advanced stages, so standards help avoid duplicate investment later. Requirements and evaluation criteria differ by program, so always check that year's official announcement.
KITIM provides data readiness assessments, supplier selection support, and business plan advisory. If you are about to order a smart factory and are unsure how far to take data standards, please reach out through KITIM's consultation request.