South Korea Restructures Artificial Intelligence Strategy Focusing on Frontier Models

South Korea's Ministry of Science and ICT is considering updating its strategy to focus on cutting-edge frontier models due to rapid changes and increasing competitive conditions in the global artificial intelligence sector.

Global Competition and Strategic Change

South Korean Minister of Science and ICT Bae Kyung-hoon and officials are reviewing the country's sovereign artificial intelligence foundation model project and other related programs.

Rapid technological developments in the sector are steadily shortening the time window for South Korea to maintain its competitiveness against leading American and Chinese systems.

International Model Developments

Recent data from the Artificial Analysis Database showed that global models such as Anthropic's Claude Opus, OpenAI's GPT, China's Moonshot AI (Kimi), and Alibaba's Qwen have recorded significant advancements.

Minister Bae's Assessments

Speaking at the Second Science, Technology, and AI Future Strategy Meeting, Minister Bae Kyung-hoon stated that the standards set 18 months ago for the sovereign AI project may no longer be sufficient.

Bae emphasized that in an industry that changes every three to six months, 2027 at the latest will be a critical threshold determining the outcome of the artificial intelligence race.

Funding Needs and Budget Studies

In his statement in July, Minister Bae noted that developing a model at Anthropic's Mythos level could require 10,000 Nvidia Vera Rubin graphics processing units and a cost of approximately 3.5 trillion won ($2.55 billion).

The ministry continues discussions with budget authorities and relevant government agencies on a new support mechanism.