US Suddenly Controls Export of Advanced AI 'Mythos', Why Samsung and SK are on High Alert
The U.S. government has restricted exports of Anthropic's latest AI model, 'Mythos 5,' citing security concerns, raising red flags for the AI strategies of Samsung and SK.

The U.S. government has unexpectedly implemented strict export controls on Anthropic's top-tier AI models, 'Mythos 5' and 'Fable 5,' citing cybersecurity and national security concerns. Consequently, major South Korean tech giants like Samsung Electronics and SK Group, who had just joined a collaborative research initiative for these models, have seen their AI strategies hit a major roadblock.
U.S. Government Extends Control to Complete AI Models
The primary catalyst for this export control is the 'Jailbreak' risk associated with generative AI. Following the discovery of critical security vulnerabilities that could be exploited for cyber terrorism or designing biochemical weapons, the U.S. government immediately classified the Mythos and Fable models as national security assets. This is highly disruptive to the market, as it marks a significant shift: export restrictions, previously limited to semiconductor chips or manufacturing equipment, have now expanded to encompass completed software products, specifically Large Language Models (LLMs). Access has been entirely blocked not only outside the U.S. but also for foreign nationals residing in the U.S., including foreign employees working at Anthropic.
Samsung and SK Hit Hard Shortly After Joining 'Project Glasswing'
Among South Korean companies, Samsung Electronics, SK Hynix, and SK Telecom are facing the most direct impact. They recently joined 'Project Glasswing,' a global AI security collaboration led by Anthropic, successfully securing access to the Mythos models. However, losing this access just ten days after joining means an inevitable and complete revision of their roadmaps for technology optimization and joint vulnerability research using top-performing AI. This serves as a harsh blow to domestic companies heavily reliant on AI models from overseas Big Tech.
The Urgency of 'Sovereign AI' for Technological Independence
This incident has proven that global AI models can be weaponized for diplomatic and security purposes at any moment. As a result, securing 'Sovereign AI' capabilities—building independent systems free from the control of specific nations—is rapidly emerging as a survival imperative for both nations and corporations. The South Korean government, including the Ministry of Science and ICT, has mobilized a consultative body centered around the National Security Office to assess the situation accurately and formulate countermeasures.
FAQ: Key Questions on the U.S. AI Export Controls
-
Q. Will the Mythos 5 export control affect everyday users?
A. While there won't be immediate changes to existing services for general users, global companies are highly likely to face delays in launching new services or devices (like on-device AI) that utilize these latest AI models. -
Q. How is this affecting the stock prices of semiconductor companies?
A. Leading South Korean semiconductor companies, such as Samsung Electronics and SK Hynix, are experiencing short-term downward pressure and dampened investor sentiment due to concerns over disruptions to their AI roadmaps. -
Q. How should domestic companies respond going forward?
A. In the short term, they must accelerate the development of their own foundational models. Long-term, they need to quickly build an independent Sovereign AI ecosystem by collaborating with open-source AI alliances outside the U.S.