AI 데이터센터 확산 정책의 통상법적 쟁점에 관한 연구

As AI technologies expand and computing demand grows, AI data centers have become a strategic infrastructure that many countries seek to secure. Their high energy consumption, concentrated GPU supply chains, and security-related risks distinguish them from conventional cloud data centers and have led governments to adopt more active support and location-focused policies. Major economies have already widened such interventions. A comparative review of these policies indicates that many AI data center support measures raise concerns about their consistency with WTO rules, particularly regarding non-discrimination, market access, subsidies, and government procurement. Measures based on security concerns, including restrictions on certain foreign technologies or domestic technology requirements, appear particularly vulnerable under existing WTO jurisprudence. An assessment of the GATT Article XXI security exception further suggests that these measures are unlikely to satisfy the high threshold applied in recent cases, which restricts “emergency in international relations” to situations approaching armed conflict. Overall, the analysis indicates that national AI data center strategies will introduce new legal and institutional challenges for multilateral and regional trade regimes. To address these developments, it will be necessary to ensure that the security exception is not applied in an overly expansive manner and to broaden the scope of regulatory discussions from AI technologies to AI infrastructure.