[Edaily Reporter Kim Hyun-ah ] Anthropic’s recent decision to restrict foreign access—a move that has shaken the global AI market—is not merely a passing incident. It is the result of a convergence between the power struggle among U.S. AI companies and the vested interests of Washington’s political circles. However, there is a more fundamental issue that deserves attention.

It has been proven that even the closest ally can pull the plug on high-performance AI at any time if it serves its national interests and security—a harsh reality. In an era where AI has become the core infrastructure driving national systems—including defense, finance, telecommunications, and manufacturing—blind reliance on the technology of a specific country poses a fatal security threat.

What is needed now is not to ask, “Which telecom company linked to China was the problem?” but to find an answer to the question: “If the U.S. were to suddenly shut off the AI switch one day, how would South Korea survive?”

AI has become a matter of geopolitics, not just technology

. Experts did not view this incident as a simple issue of technology regulation. Many interpreted it as the result of a complex interplay of competition among U.S. AI companies, policy decisions by the White House, and the national security framework.

In fact, U.S. policymakers no longer view AI as a private-sector service but as a strategic national asset. There is a growing recognition that it is critical infrastructure—like semiconductors, energy, and telecommunications networks—that determines a nation’s competitiveness.

One expert stated, “The core of this controversy is not a specific South Korean company,” adding, “The fact itself—that the U.S. can restrict AI access even to its allies if it deems it necessary—is what matters.”

This is where the problem lies. No matter how diligently South Korean companies operate their services, ultimate control rests with the country that owns the Frontier AI model. This is why AI sovereignty has become a practical national challenge rather than an abstract slogan.

Just as the semiconductor supply chain has become a matter of national security, AI models and computing infrastructure are also emerging as new strategic assets. AI is no longer confined to the realm of technological competition; it has entered the realm of geopolitics.

What is even more painful is the lack of preparedness among Korean companies

. However, we cannot simply shift all the blame onto the United States. It is regrettable that Korea was ill-prepared, even though it had ample time to prepare for this situation.

SamsungElectronics(005930)Leading domestic companies such as Naver (NAVER(035420)) have all experienced internal bottlenecks during their AI transitions.

In the case of SamsungElectronics, while the company recruited key talent from overseas Big Tech firms, assessments suggest that their capabilities were not fully utilized due to organizational barriers and a business-unit-centric culture. Some have also pointed out that resources were scattered because similar research projects were being pursued redundantly across multiple organizations.

The shortcomings surrounding Naver are even more specific. Although this happened in the past, it is said that there was such a skeptical view of AI investment that, during the approval process for investments in super-large AI infrastructure, executives reportedly made comments to the effect of, “Please play well with this 1200억-won toy.”

The problem wasn’t just about money. It was about differences in decision-making speed, organizational culture, and vision for the future.

While OpenAI, Google, and Meta poured tens of trillions of won into AI and focused their entire corporate resources on it, Korean companies remained relatively cautious. As a result, the gap widened faster than expected.

This is why the public-private partnership known as “Sovereign AI” is so important

. For this reason, the “Frontier AI” strategy recently unveiled by the government carries significance beyond that of a simple industrial policy.

Minister Bae Kyung-hoon of the Ministry of Science and ICT presented a vision to move beyond AI specialized for manufacturing and take on the challenge of developing world-class general-purpose AI models (Frontier AI). Expanding GPU infrastructure, fostering domestically produced AI semiconductors, and building a data ecosystem are all part of this same effort.

Of course, it is difficult for South Korea to compete on a capital scale with U.S. Big Tech. However, “Sovereign AI” does not mean developing everything independently, nor is it something a single company can achieve on its own.

The key is to secure a minimum level of control over core models, data, and computing infrastructure, and to create a structure that ensures the national economy and industries do not grind to a halt even in the event of external shocks.

Just as South Korea did in the past with the semiconductor industry and the mobile communications sector—including TDX (domestically produced electronic switching systems) and CDMA (Code Division Multiple Access)—the public and private sectors must join forces to make long-term investments and build an AI ecosystem. The government, businesses, and academia must share the risks and work together to grow the market.

AI sovereignty is not achieved through mere declarations. It becomes possible only when infrastructure, talent, data, and markets are accumulated together.

Technological sovereignty is not a choice but a matter of survival

. The recent Anthropic controversy has served as an important wake-up call for South Korea.

AI is no longer merely a convenient service. It is a strategic asset that determines national security and industrial competitiveness. Whether it is the United States or China, the supply of technology can become a weapon of diplomatic and industrial policy at any moment when national interests are at stake.

If the South Korean economy is to keep running even on the day the U.S. flips the AI switch off, we must begin preparing for sovereign AI and AI infrastructure sovereignty at the national strategic level starting now. Technological sovereignty is no longer merely a matter of industrial policy. It is a matter of survival.