[Edaily Reporter Kwon Oh Seok ] As investment in artificial intelligence (AI) data centers expands, market interest is broadening beyond power equipment to include companies involved in data center construction methods. In particular, modular data centers—in which key structural components, power systems, and cooling equipment are prefabricated in factories and assembled on-site—are gaining attention as an alternative that can shorten construction time.

According to MP Doctor on the 7th, iA, Inc.(038880)was trading at 3,240 won as of 11:10 a.m. that day, up 9.64% from the previous trading day. During the session, the stock rose more than 11%, showing strong momentum. In addition, other stocks related to AI data center power infrastructure—such as HD HYUNDAI ELECTRIC(267260), LS ELECTRIC(010120), HYOSUNG HEAVY INDUSTRIES(298040), SANIL ELECTRIC, and DOOSAN ENERBILITY(034020) —are also showing an upward trend.
As the AI data center market expands, the scope of investment is broadening to encompass the entire infrastructure, including power generation, transmission and distribution, cooling, and data center construction. In a recent analysis of the power equipment industry, Yuanta Securities Korea noted that new orders this year for HYOSUNG HEAVY INDUSTRIES, HD HYUNDAI ELECTRIC, LS ELECTRIC, and SANIL ELECTRIC are likely to exceed the companies’ existing guidance by approximately 20% based on the median estimate.
At the same time, the “modularization” of data centers has recently emerged as a new investment keyword. This is because, as demand for AI infrastructure grows rapidly, it is becoming increasingly important not only to build data centers on a large scale but also to be able to construct and operate them quickly when needed.
LG Uplus is developing a standard model for prefabricated modular data centers (PMDC) in collaboration with GS Engineering & Construction Corp., Zai C&A, Gansam Architects & Partners, and D&O CM. The company is promoting the standardization of data center construction by reviewing models ranging from those applicable to 100MW-class hyperscale data centers to retrofit models that convert existing buildings into data centers.
LS ELECTRIC also signed a strategic partnership with KTCorporation Cloud on the 2nd to develop AI data centers. The two companies agreed to supply power equipment—including ultra-high-voltage transformers, GIS, and power distribution panels—and to conduct joint research on modular data centers. With major data center operators and power equipment manufacturers now participating in the construction of modular data centers, the related market is likely to expand significantly.
There are forecasts that modularization can expand its scope beyond large-scale hyperscale data centers to include small-scale, distributed infrastructure at the site level that requires AI computation. This is because, as demand for AI inference expands into various sectors such as manufacturing, defense, and the public sector, there may be a need to build AI computing infrastructure directly on-site, separate from centralized data centers.
In response to these market changes, iA, Inc. is promoting the “NeoCube” project, a modular AI data center model with a capacity of 1 MW or less. NeoCube is designed to perform AI computations at the site level by integrating power, cooling, and cloud infrastructure. The company views sectors where data security and low-latency processing are critical—such as manufacturing, defense, the public sector, finance, and healthcare—as key target markets.
At last month’s OCP Korea Tech Day, iA, Inc. highlighted the potential for AI inference infrastructure to expand to regional and on-site levels amid the proliferation of sovereign AI and physical AI, and unveiled NeoCube as a reference design.
The industry is focusing on the potential for modular data centers to grow into a market that complements the demand for distributed AI infrastructure and accelerates deployment, rather than replacing existing hyperscale data centers. However, some observers point out that for this to lead to an actual revaluation of the company’s market value, it will be necessary to confirm not only the technology and market outlook but also deployment cases, commercial contracts, and revenue contributions.