SK Telecom Triggered Anthropic Mythos Export Controls; White House Revoked Access Over China Ties

The Trump administration ordered Anthropic to revoke SK Telecom's access to Claude Mythos on June 12, 2026, after identifying the South Korean carrier as having suspected ties to China, according to reporting by The Washington Post and WIRED. SK Telecom, South Korea's largest wireless carrier and a $100 million Anthropic investor since 2023, had been granted access to Mythos through Anthropic's Project Glasswing on June 4. The White House identified SK Telecom on Anthropic's expanded access list and demanded the company revoke the carrier's credentials immediately.

SK Telecom's footprint in China is minimal: roughly $1.9 million in Chinese revenue in 2024 with seven employees. However, its parent SK Group holds extensive interests in Chinese semiconductors and energy. SK Telecom formed a joint venture with China Unicom in 2004, invested $1 billion in Unicom bonds in 2006 (converting to 6.6% stake, later sold in 2009 for $1.3 billion), and retains a UNISK investment worth ~$17 million as of 2025. This history triggered White House skepticism about Anthropic's partner vetting.

Anthropic complied immediately, revoking SK Telecom's access the same day. Days later, Amazon (Anthropic's largest investor with ~$13 billion cumulative stake) flagged potential guardrail bypasses in Fable 5 to the Commerce Department, leading to broader export controls on June 12 barring all foreign nationals—including foreign Anthropic employees—from accessing both Fable 5 and Mythos 5. Rather than implement nationality-based access controls, Anthropic disabled both models entirely for all users.

SK Telecom denied allegations, telling Korean media it has no China ties and does not use Huawei in core networks. On June 17, Anthropic opened a Seoul office and signed an MOU with South Korea's Ministry of Science and ICT, naming SK Telecom as a local partner. Negotiations between Anthropic and the White House continue; Anthropic's international managing director said at the Seoul launch he was very confident both models would return in the coming days. About 100 cybersecurity professionals, including ex-Facebook security chief Alex Stamos, signed a letter arguing Mythos-class models are not uniquely good at finding vulnerabilities and called for the controls to be lifted.

This episode illustrates how geopolitical vetting and export controls can disrupt AI model access even for established partners in allied nations. Architects deploying advanced AI models across international teams should expect increased due diligence on corporate structure, investment history, and supply-chain exposure to sanctioned jurisdictions. The case also shows the technical challenge of implementing nationality-based access controls—Anthropic found it simpler to disable models entirely than to verify user nationality.