South Korea is restructuring its Dokpamo independent AI foundation model initiative after broader ecosystem assessments placed the country third globally behind the United States and China, while a newer Artificial Analysis evaluation showed no South Korean model in the global top 25. The government plans to launch a separate Frontier AI project next year for top-tier model development and refocus Dokpamo on deploying domestic AI through industry and public services. It plans to allocate 5.5937 trillion won, about $4.1 billion, for next year’s science and ICT budget, including roughly 4.7 trillion won, about $3.4 billion, for a Frontier AI company formed by a private consortium. Dokpamo is expected to be folded into the Frontier project after its third-round evaluation, which runs through early next year, although the governance structure and operator-selection process remain undecided. The AI for All program, led by Kakao, SK Telecom and KT, is intended to bring domestic models into general-purpose chatbots, public AI agents and specialized services, with a beta launch planned for late September and a full launch in December. Motif 3 was South Korea’s highest-ranked model in the newer AAII assessment, placing 36th with 33.6 points, or 63% of the leading model’s score; all 25 top positions went to U.S. or Chinese models. The strategy faces criticism over fragmented project awards, evaluation fairness, limited computing resources and a large research and investment gap, while officials debate how to accelerate development alongside safety, cybersecurity and resource-efficiency safeguards. The government’s approximately 9.4 trillion-won AI budget proposal, including Frontier AI spending, remains subject to National Assembly review.