South Korea is making a bold move in the global AI race. It plans to construct one of the world’s largest AI datacenters in Jeollanam-do, the province at the southern tip of the Korean Peninsula, with a planned capacity of 3 gigawatts (GW) – nearly three times that of the planned initial site for Stargate, the AI datacenter joint venture announced recently by U.S. President Donald Trump. At a cost of $35 billion, the initiative is more than just an infrastructure project. It is one of several government initiatives that represent South Korea’s leap from a “fast follower” to an AI leader with a robust, homegrown ecosystem.
However, as AI becomes politicized and increasingly used as a medium for strategic competition, South Korea must chart a delicate course through uncertain waters. The country finds itself in a uniquely precarious position as it seeks technological AI sovereignty to secure its own interests, while attempting to balance deep economic ties with China and an existential security alliance with the United States.
South Korea’s AI Sovereignty Push
South Korea has long been a global leader in the semiconductor industry, particularly in high-bandwidth memory (HBM) design and chip fabrication, two critical pillars of the AI supply chain. Samsung and SK Hynix together control approximately 90 percent of the HBM market, and their development of 12-layer HBM3E chips ensures a stable supply of high-performance components for AI accelerators. HBM technology and microchip fabrication is essential to the AI supply chain and South Korea’s dominance in these markets strengthens its AI development prospects.
Despite its leadership in these fields, South Korea’s role in other areas of the AI value chain remains limited. It continues to heavily rely on foreign hyperscalers for its cloud computing needs, lacks globally competitive foundation models, and has historically lagged behind other countries on the development of AI-based applications. This contrast highlights the critical gap South Korea must bridge to establish technological sovereignty.
In response, the government has launched a series of initiatives to spur domestic AI innovation. At the third National AI Committee meeting in February, it unveiled the “World’s Best LLM” project – an effort designed to attract, incentivize, and support top-tier AI teams with large-scale computational resources and substantial research funding, including a $750 million public investment to advance general AI technologies. Additional measures include enhanced support for emerging domestic talent and the recruitment of leading international experts to work and research in South Korea. To facilitate domestic AI research, the government has also laid out plans to provide AI researchers access to high-quality data for AI training, including special provisions that would enable the use of certain types of personal data in a secure manner.
The planned datacenter in Jeollanam-do is a natural extension of these efforts. It will be a cornerstone of South Korea’s AI strategy, serving as the computational backbone of the broader effort to build a closed-loop AI ecosystem. In this model, domestically produced semiconductors would power domestic datacenters, including the facility in Jeollanam-do, which would in turn support the development and deployment of domestic large language models (LLMs) and their AI applications. As adoption expands, this cycle will become self-reinforcing, generating value at every step that can in turn be reinvested into next-generation semiconductor and AI research.
The goal is for the closed-loop system to drive the entire South Korean AI economy from a position of security and allow South Korea to reap the economic benefits of AI, while positioning itself as a global AI leader. While the Jeollanam-do datacenter is only a single part of a much broader set of industrial policies aimed at promoting AI development, it is the most visible symbol of South Korea’s growing ambition and a manifestation of the government’s desire to secure both its self-sufficiency and its foothold within the new AI-powered industrial revolution.
South Korea’s Ambitions and China-U.S. Competition
South Korea’s pursuit of AI sovereignty is largely driven by its desire for strategic autonomy and economic growth. Seoul is aiming to develop AI systems that are not controlled by foreign entities. Sovereignty is crucial for maintaining control over key areas of its economy and ensuring that AI technologies align with South Korea’s political values and priorities, including the need to maintain data privacy and security.
As a technology exporter, South Korea additionally recognizes that AI represents a possible multi-trillion-dollar opportunity to foster economic growth beyond its traditional industries. Furthermore, applications of AI to areas of automation and manufacturing offer a potential lifeline for South Korea’s impending labor shortage, a consequence of having the lowest birthrate in the world. In all facets, it is clear that developing domestic AI capabilities is imperative, and that these will serve as a strategic resource that will enable Seoul to cement its position as a leader in emerging technologies and establish its plan for long-term strategic resilience.
Yet South Korea currently finds itself caught within the escalating China-U.S. global competition. Economically tied to China, its largest trading partner, it also has a decades-long defense agreement with the U.S., which dates back to the Korean War. While Seoul is no stranger to this delicate balancing act, recent escalations in the China-U.S. strategic rivalry have substantially increased the pressure that South Korea faces on the geopolitical stage. The growing polarization between the U.S. and China is particularly evident in the race to lead on AI, resulting in U.S. export restrictions on advanced semiconductors to China and the geopolitically charged media coverage of DeepSeek, a Chinese LLM. Amidst this backdrop, Seoul finds its actions being scrutinized by both sides, adding considerable complexity to South Korea’s efforts to assert AI sovereignty.
At the time of writing, the U.S. has extended an invitation to the country to join the Stargate project, aiming to leverage South Korea’s (and specifically Samsung’s) dominance in HBM design and semiconductor fabrication to build a security-first AI framework aimed at developing cutting-edge AI infrastructure within the U.S. Accepting the offer comes with several benefits. It would position South Korea as a key partner in Stargate’s AI model development, help strengthen its bargaining power with the Trump administration, and provide the country with access to cutting-edge AI advancements.
However, integration with the U.S.-led ecosystem comes at the cost of potentially jeopardizing South Korea’s AI self-reliance. South Koreans are uneasy with the U.S. government’s nationalistic AI policies, including the expected implementation of the Stargate project, which are perceived to heavily favor U.S. companies at the expense of foreign partners. Based on a 2023 poll of the general public, 77 percent of Koreans believed that the CHIPS and Science Act, signed into law in 2022 to boost the U.S. semiconductor industry, was driven by “America first” principles. and 82.6 percent believed the U.S. needs to consider its allies’ interests more carefully. These insights highlight a growing concern among the public regarding South Korea’s subordinate position in its dealings with the United States.
Alignment with the U.S. also risks provoking retaliation from China, as has been seen during past instances of geopolitical friction. Recent tensions, such as South Korea’s ban of DeepSeek over data security and historical disputes, underscore how fragmented the China-South Korea relationship remains. When South Korea deployed the Terminal High Altitude Area Defense missile defense system in 2017, China responded with sweeping economic sanctions targeting the Korean entertainment, tourism, and consumer goods industries. A similar backlash could just as easily emerge in response to a South Korean decision to cooperate with U.S. AI policy. China has only expanded its toolkit for punishing foreign companies in the years since.
The semiconductor industry, a critical component of the AI supply chain, is particularly vulnerable, with 33.3 percent of South Korea’s chip exports going to China. Moreover, Korean chipmakers have grown increasingly dependent on Chinese raw materials, including rare earth minerals – a trend that shows no signs of slowing. In just one year from 2022 to 2023, South Korea imports of Chinese silicon increased from 68.8 percent of total supply to 75.4 percent, with the total volume of imports also growing significantly. As such, it is clear that while participation in initiatives like Stargate may serve U.S. strategic interests, it also carries significant economic and geopolitical risks for South Korea.
On the other hand, if South Korea pushes too aggressively toward AI independence, it risks prompting U.S. retaliation, possibly in the form of tariffs on Korean exports. South Korea’s advancements in AI technologies may eventually challenge U.S. technological dominance while reducing bilateral collaboration, likely exacerbating trade imbalances. Such a perceived asymmetry could prompt Washington to impose tariffs to reduce its trade deficit and protect its own AI ecosystem. The risk is not hypothetical: in April, Trump announced tariffs on nearly every country in a bid to reduce the U.S. trade deficit.
While chip imports were given a temporary reprieve, Trump has indicated that tariffs on foreign semiconductor products will follow soon. In response, just two days later, South Korea introduced the enactment of a $23 billion financial support package for the semiconductor industry amid growing uncertainty over future trade conditions. Through this package, the government is providing various financial incentives to boost growth in the semiconductor industries, such as low-interest loans, subsidies, and the acquisition of cutting-edge manufacturing equipment.
Notably, the temporary reprieve for semiconductor imports suggests that Washington considers chipmaking to be both a strategic asset and a key vulnerability. This risk has triggered mounting concern in South Korea that chips could become the next target in Washington’s push for onshoring. This would place Korean chipmakers in a difficult position as they face pressure to expand manufacturing to the U.S., potentially exposing them to financial and security risks.
Mapping Out South Korea’s Future AI Trajectory
The Jeollanam-do datacenter is significantly more than a technological milestone; it is a symbol of Seoul’s determination to control its own AI future. It is clear that South Korea is no longer content to remain a passive player in the AI revolution. However, the true challenge lies not in building infrastructure or providing funding, but in converting these efforts into sustainable AI leadership without compromising key international relationships.
As the U.S. and China continue to push forward with their AI initiatives and advance their geopolitical competition, South Korea must find a way to navigate the fine line between strategic autonomy and global interdependence. Leveraging its existing strength in semiconductors and utilizing sustained public investment, Seoul must ensure that its pursuit of technological sovereignty encourages the development of a globally competitive AI industry, while preserving its own national security and economic interests.
The geopolitical tightrope must be traversed carefully, but the transformative potential of AI and the critical importance of sovereignty are far too great to ignore.