The rapid global increase in data, coupled with advancements in artificial intelligence (AI) technology, has made data center construction an essential task worldwide. However, news that even the United States, known as an IT powerhouse, is facing severe difficulties in data center construction may come as a shock. Let's examine how this issue might impact the global IT market and what implications it holds for South Korea. According to a Bloomberg report, nearly half of the US data center construction projects planned until 2026 are at risk of delay or cancellation. This is primarily due to shortages in essential power infrastructure and critical electrical components required for data center operations. Data from market intelligence firm Sightline Climate indicates that while approximately 12 gigawatts (GW) of data center capacity were expected to be operational in the US by 2026, only about one-third of that capacity is currently under active construction. Notably, despite tech giants like Alphabet (Google's parent company), Amazon, Meta, and Microsoft announcing plans to invest over $650 billion in AI infrastructure expansion, power-related bottlenecks, rather than funding or computing technology, are hindering their progress. The biggest problem is the woefully insufficient production of critical components such as transformers, switchgear, and batteries compared to demand. These pieces of equipment are essential not only for the data centers themselves but also for the power grid infrastructure that AI companies need to expand to ensure adequate power supply. Benjamin Boucher, a senior analyst at Wood Mackenzie, pointed out that domestic production capacity cannot keep pace with demand, forcing reliance on export markets. Although power infrastructure accounts for less than 10% of data center construction costs, the inability to secure it can halt an entire project, posing a serious challenge. The fact that US transformer manufacturers largely depend on foreign imports due to their inability to expand domestic production capacity further highlights these structural limitations. **Problems Caused by US Supply Chain Bottlenecks** The increasing demand for energy-efficient products like electric vehicles and heat pumps exacerbates the problem. These products also require the same core electrical components as data centers, leading to competition for limited supplies within the supply chain. For instance, transformer delivery times, which once ranged from 2 to 2.5 years, have now stretched to up to 5 years. This far exceeds the needs of AI companies and poses a significant obstacle not only to data center construction plans but also to power grid upgrades across the US. Some experts warn that this issue, given that the entire world faces similar component supply chain problems, will not be resolved in the short term. Considering that the ability to process data is central to corporate competitiveness, and power infrastructure is becoming a critical bottleneck, this issue is expected to have a substantial impact on the global IT industry, not just the US. While several alternatives are being discussed to address these problems, the practical constraints are considerable. Some data center developers are attempting to temporarily resolve the issue by pre-ordering equipment in bulk or recycling old transformers from decommissioned power plants. However, these solutions cannot fully alleviate market uncertainty as long as the fundamental problem remains unresolved. Indeed, these delays are expected to worsen beyond 2027, with many projects yet to even break ground. Despite ongoing efforts by major data center companies to adopt sustainable energy solutions and improve energy efficiency, a rapid structural transformation will be necessary to meet the demands of AI-driven expansion. Furthermore, there is a growing societal demand for sustainable alternatives, such as improving data center energy efficiency and transitioning to clean energy. At this point, some counterarguments are also anticipated. Some argue that this crisis is a short-term problem that will naturally resolve itself with the pace of AI and data processing technology development. However, given that the root cause of the bottleneck lies in manufacturing capacity and supply chain limitations, it is highly probable that technological advancements alone will not suffice. For example, expanding US domestic production capacity requires time-consuming efforts such as building new manufacturing facilities, deploying personnel, and strengthening logistics systems. With US manufacturers currently unable to keep pace with production, it is undeniable that this issue will have ripple effects across the entire global supply chain. **Korea's Potential in the Global Data Center Market** So, what opportunities can South Korea seize? Although the original source material for this analysis does not directly mention Korea, several implications can be drawn consid
Related Articles