[Kigyeong Hyeok AX Column] The Super Cycle of Medium-Voltage Equipment Now Begins

[ZDNet Korea] Global AI data center investment is surging, coinciding with demand for replacing aging power grids, propelling the medium-voltage equipment market—comprising high-voltage transformers, circuit breakers, and GIS (Gas-Insulated Switchgear)—into a historic supercycle. Domestic giants like Hyosung Heavy Industries, LS Electric, and HD Hyundai Electric face backlogs exceeding several years, with stock prices and results hitting unprecedented highs. However, beneath this prosperity lies a paradox: soaring orders clash with delayed deliveries. Even with expanded production, output lags—because the core issue is skilled labor, not equipment. Unlike semiconductor or display industries where machinery automates processes, medium-voltage equipment relies on decades of field expertise in insulation design, arc interruption judgment, and assembly/testing. This knowledge isn't transferable via manuals but is instead internalized through hands-on experience. The crisis stems from insufficient attention to this sector compared to semiconductors, which enjoy national talent acquisition priorities and top-tier compensation. Meanwhile, medium-voltage equipment faces talent drain to semiconductor and IT firms, exacerbating labor shortages. The supercycle’s success hinges on workforce availability, as lacking domain expertise creates structural limits to production scaling. Here, AX (AI Transformation) redefines its role: it’s not about adopting generic AI tools but converting engineers’ judgment logic, operational experience, and process know-how into rule-based data models (digital specifications). Modeling an expert’s knowledge allows it to operate across multiple sites simultaneously. Yet, a critical prerequisite exists: explainable AI (XAI) must validate and articulate why algorithms make specific decisions to earn global trust. Beyond productivity gains, AX enables unique data assets—real-time operational data from globally deployed equipment (e.g., current/voltage patterns, anomaly indicators, load variation histories)—which become irreplaceable intellectual property. These data underpin predictive maintenance services and…