What Happened to SK Telecom's Anthropic Mythos Controversy?

The "SK Telecom's Anthropic Mythos Controversy" emerged in June 2026, revealing the escalating geopolitical complexities surrounding advanced artificial intelligence development and international partnerships. The core of the dispute centers on the US government's intervention regarding South Korean telecommunications giant SK Telecom's access to Anthropic's highly advanced Claude Mythos AI model.

The relationship between SK Telecom and Anthropic began in August 2023 when SKT made a strategic investment of $100 million in the US-based AI safety and research company. This investment was coupled with a partnership agreement to jointly develop a multilingual large language model (LLM) specifically customized for telco use cases, leveraging Anthropic's Claude models. SKT also became the only Asian telecom company to participate in Anthropic's security collaboration initiative, "Project Glasswing".

The controversy ignited in early June 2026 when Anthropic expanded its Project Glasswing program, granting approximately 150 organizations across more than 15 countries, including SK Telecom, Samsung Electronics, and the Korea Internet & Security Agency, access to its specialized cybersecurity AI model, Claude Mythos. However, the White House quickly raised concerns over what it described as SK Telecom's alleged ties to China. While SK Telecom itself has a minimal direct footprint in China, it is part of the broader SK Group, which possesses extensive Chinese interests in energy and semiconductors. Furthermore, SK Telecom had a past joint venture with state-owned China Unicom and maintained a financial interest in it as of a 2025 SEC filing.

Days before Anthropic publicly announced the US government's restrictions, the White House reportedly ordered Anthropic to revoke SK Telecom's access to Claude Mythos. Anthropic immediately complied with this directive. This incident, combined with separate warnings from Amazon researchers who reportedly found ways to "jailbreak" Anthropic's public Fable 5 model (a version of Mythos) and bypass its safety guardrails, led the US government to conclude it could not trust Anthropic to adequately safeguard its cutting-edge technology.

Consequently, on June 12, 2026, the Trump administration issued an order restricting access to Mythos 5 and Fable 5 for all foreign nationals under national security authorities. Rather than implementing a complex nationality-based gatekeeping system, Anthropic chose to take both models entirely offline, affecting foreign nationals both inside and outside the United States. This decision left Anthropic and the White House in a deadlock after tense negotiations.

Despite these developments, SK Telecom's CEO, Jung Jai-hun, announced on June 10, 2026, that the company had made an additional, undisclosed investment in Anthropic. Jung stated that the purpose was to maintain a cooperative relationship rather than focusing on potential IPO profits, emphasizing ongoing collaboration in areas like infrastructure, data, and security. SK Telecom subsequently denied any alleged ties to China, stating that the allegations lacked verified facts. As of June 19, 2026, the controversy underscores the increasing scrutiny of international AI partnerships, particularly those involving companies with perceived links to geopolitical rivals, and highlights the challenges of balancing global innovation with national security concerns in the rapidly evolving AI landscape.