The South Korean government, which has recently focused on supporting domestic neural processing units (NPUs), is now turning its attention to data centers. It is broadening its localization support from servers and storage to networking, power and cooling equipment, and operational software (SW). The move is interpreted as a preemptive response to repeated equipment delivery delays caused by global supply chain instability at a time when the data center market is expanding rapidly due to growing artificial intelligence (AI) adoption.
According to science and technology sector sources on Wednesday, the Ministry of Science and ICT (MSIT) recently issued a consolidated call for its "2026 Sustainable Data Center Industry Development Support Program." The initiative covers development and field testing of domestic equipment and software, development and field testing of eco-friendly and high-efficiency equipment and software, an industry survey, and an awareness improvement campaign. Aimed at building a foundation for sustainable growth of data centers — a core infrastructure of the AI era — the program supports field testing of equipment and solutions developed by small and mid-sized Korean companies.
The government's direct involvement in field testing of data center equipment stems from an urgency to quickly channel surging AI infrastructure investment demand into the domestic industrial ecosystem. Large-scale computing resources for AI training and inference, massive data processing, and cloud service expansion are all advancing simultaneously, driving a sharp increase in data center demand. According to market research firm Spherical Insights & Consulting, the global AI data center market is projected to grow roughly tenfold, from $1.28 billion in 2023 to $12.581 billion in 2033.
AI semiconductors are the most critical component in building a data center, but equipment such as power, cooling, and networking systems is also essential. A delay in any single element of the power supply chain can halt an entire project. Yet surging data center demand is lengthening equipment lead times worldwide. Bloomberg recently cited Sightline Climate as reporting that lead times for high-capacity transformers in the United States, typically 24 to 30 months before 2020, could now stretch to as long as five years. Given that AI data centers generally assume a construction cycle of around 18 months, such delays can be critical.
In Korea's case, a significant share of core equipment installed in domestic data centers currently depends on foreign suppliers, raising substantial supply chain concerns. According to MSIT, estimated domestic equipment utilization rates as of 2024 stand at 3.3% for servers, 1.82% for storage, 23.38% for uninterruptible power supplies (UPS), and 75.83% for batteries. Servers and storage — the key IT equipment for AI data centers — remain overwhelmingly dependent on foreign products. Under this structure, any disruption in the imported equipment supply chain could lead not only to cost increases but also to delays in AI data center construction.
"If servers, networking, cooling, power, and operational platforms are all foreign-made, the entire ecosystem risks becoming dependent on overseas technology," an industry official said. "For domestic AI semiconductors to be stably commercialized inside actual data centers, the surrounding equipment and operational ecosystem must be cultivated alongside them."