KISDI Recommends ‘Minimum Dependence, Maximum Competitiveness’ as Korea’s Strategy for AI Technological Sovereignty (detail)

※ URL(Korean): https://www.kisdi.re.kr/bbs/view.do?bbsSn=115006&key=m2101113055776&pageIndex=1&sc=&sw=

KISDI Basic Research (25-14)
Strategies for Strengthening AI Technological Sovereignty and National Competitiveness

KISDI Recommends ‘Minimum Dependence, Maximum Competitiveness’ as Korea’s Strategy for AI Technological Sovereignty
.................................................................................................................................................
▲ Achieving full-stack AI self-sufficiency is unrealistic; strategic investment should focus on core technology stacks
▲ In the short term, priority should be given to strengthening the AI model training ecosystem, including AI models and cloud infrastructure
▲ In the medium to long term, Korea should build an AI ecosystem linked to domestic semiconductors and system software

The Korea Information Society Development Institute (KISDI, President Sangkyu Rhee) recently published its Basic Research Report (25-14), *Strategies for Strengthening AI Technological Sovereignty and National Competitiveness*. The report presents Korea's strategic direction for securing AI technological sovereignty and outlines six key policy initiatives in response to intensifying global competition in AI.

AI has emerged not only as a tool for improving industrial productivity but also as a transformative technology shaping socioeconomic development and national security. At the same time, competition for AI leadership among major economies—including the United States, China, and the European Union—has intensified. Against this backdrop, the report analyzes global AI technology and market trends, reviews the AI ecosystems of major economies, and assesses Korea's AI technological capabilities, industrial competitiveness, and policy environment.

Based on its analysis, KISDI argues that the objective of AI technological sovereignty should not be the creation of a fully self-sufficient AI ecosystem. Instead, it recommends redefining the goal as achieving **"minimum dependence and maximum competitiveness."** Given the United States' substantial technological lead and the highly interconnected evolution of AI technology stacks, the report concludes that achieving full-stack self-sufficiency across all AI technologies is not a realistic short-term objective. Instead, Korea should concentrate its resources on technology areas where it has the greatest potential to compete internationally.

Drawing on a survey of AI technology experts, the report evaluates supply chain risks and future competitiveness across major AI technology stacks and classifies policy priorities into four categories:

1. Government-led priority investment (High supply chain risk / High competitiveness potential)

* AI foundation models
* Cloud infrastructure

2. Government-led long-term investment (High supply chain risk / Lower competitiveness potential)

* Semiconductors and AI accelerators
* System software

3. Private sector-led investment with government support (Low supply chain risk / High competitiveness potential)

* Application software
* Networks
* Data infrastructure and solutions

4. Continuous monitoring (Low supply chain risk / Lower competitiveness potential)

* Development platforms and frameworks

Overall, the report recommends prioritizing the development of a competitive AI model training ecosystem—including AI models and cloud infrastructure—in the short term. Over the medium and long term, it proposes building an AI ecosystem that can expand and interoperate based on domestically developed semiconductors and system software.

To support this strategy, the report presents six policy directions:

* Narrow the global technology gap in large-scale AI models while fostering an ecosystem in which foundation models and domain-specific models reinforce one another
* Strengthen integrated competitiveness by linking sovereign AI models with NPUs and hardware-software technologies
* Build sustainable demand for AI by expanding the AI service ecosystem
* Promote **vertical AI** and **physical AI** linked to Korea's strategic industries to accelerate industrial AI adoption
* Strengthen AI governance to balance innovation with safety
* Foster a research ecosystem that supports high-risk, high-impact research aimed at overcoming fundamental technological limitations in AI

The report also notes that major economies, including China and the European Union, are pursuing AI technological sovereignty through approaches tailored to their own circumstances. These strategies seek to strengthen linkages across AI technology stacks and secure greater control over critical segments of the AI supply chain, rather than relying on identical policy models.

Kyungsun Lee and SUNGOK KIM, Research Fellows at KISDI, emphasized that, as international experience demonstrates, achieving technological self-sufficiency across every AI technology stack is not the only path to AI sovereignty. They noted that the primary objective of AI sovereignty should be to strengthen Korea's capacity to maintain strategic autonomy and resilience in the face of external pressures.

The report is available for download from the KISDI website (http://www.kisdi.re.kr)).

KISDI News

KISDI Recommends ‘Minimum Dependence, Maximum Competitiveness’ as Korea’s Strategy for AI Technological Sovereignty