The White House Pulled the Plug on Anthropic's Two Best Models. A Korean Telecom and a Three-Word Prompt Did It.

Correction (October 9, 2026): An earlier version of this article misdescribed the 90-minute window, overstated how the shutdown was "expanded," gave David Sacks the wrong title, attributed Anthropic's position to unnamed sources, quoted security experts we could not identify, misstated the June 20 and June 22 deadlines, gave the wrong date for the Seoul office opening and the Kalshi price, and included an unverified claim about MiniMax. Those passages have been corrected or removed using the sources linked below.

On June 9, 2026, Anthropic shipped Claude Fable 5 — its first public Mythos-class model — to the Claude API and its paid plans. Three days later, the US Commerce Department, in a letter reportedly signed by Commerce Secretary Howard Lutnick, issued an export-control directive ordering Anthropic to suspend access to Fable 5 and its companion model Mythos 5 for any foreign national, inside or outside the United States. Because Anthropic could not verify nationality in real time, it did the only thing it could: it shut both models down for all customers. Seven days later, they are still offline.

How a Korean Telecom Triggered a Federal Shutdown

One trigger is Project Glasswing, Anthropic's invite-only cybersecurity consortium. Launched in April 2026 with roughly 50 partner organizations, Glasswing expanded in early June by about 150 organizations across more than 15 countries. Among the newly admitted participants, according to WIRED's reporting, was SK Telecom, South Korea's largest mobile carrier and a $100 million Anthropic investor since 2023. US officials grew alarmed at what they saw as SK Telecom's alleged ties to China, and the White House told Anthropic to cut off its access, which the company did immediately. SK Telecom has denied any China connections.

The SK Telecom episode did not happen alone. Around the same time, Amazon and other companies flagged a way around Fable 5's safeguards, and according to that reporting the two issues together cost Anthropic the White House's confidence. The Financial Times reported that Anthropic was given 90 minutes to address the issues before the export control was ordered. The reported bypass was strikingly simple: per Katie Moussouris of Luta Security, who says she was the only outside expert to read the underlying research, Fable 5 refused to review vulnerable code for security issues but complied when asked to "fix this code", and Moussouris argues it was not a jailbreak at all. The directive itself, Anthropic says, did not explain the government's specific national security concern.

The Standoff Inside Anthropic

David Sacks, co-chair of the President's Council of Advisors on Science and Technology, said publicly that the administration asked Dario Amodei to either fix the jailbreak or de-deploy the model, and that Amodei refused. Anthropic's public position is that the demonstration it reviewed surfaced only a small number of previously known, minor vulnerabilities that other public models can find without any bypass, and that a narrow potential jailbreak should not be cause for recalling a commercial model. It went further: "We suspect that perfect jailbreak resistance is not currently possible for any model provider," the company wrote. Moussouris and more than 100 cybersecurity leaders signed an open letter urging the administration to reverse the restrictions.

The impasse is consequential. Anthropic is now the world's most valuable private AI company — $965 billion after its $65 billion Series H announced in late May — with its two flagship models suspended and no confirmed timeline for restoration. The company marked the opening of its Seoul office on June 17 anyway, with Chris Ciauri, its managing director of international, telling reporters: "We are very confident that in the coming days, the models will become available again." Prediction market Kalshi priced the probability of restoration before July 1 at 57% as of June 17.

The Collateral Damage Is Already Running

While the policy dispute plays out in Washington, the downstream effects are accumulating fast. June 20 is the deadline for subscribers who paid for Fable 5 access between June 9 and June 14 to cancel for a prorated refund. June 22 is the last day of the launch window in which Fable 5 was included in Pro, Max, Team, and seat-based Enterprise plans at no extra cost. Korean enterprise customers who had just been announced alongside the Seoul office — NAVER deploying Claude Code across its engineering organization, Samsung SDS, LG CNS, Nexon, Hanwha Solutions — find themselves without Anthropic's newest model not because of anything they did, but because a telecom company in the same country made the government nervous. Anthropic's other models remain available, but the timing is hard to miss: the company's big Korea push landed the same week the models at the top of its lineup were suspended.

When a government can switch off a product overnight, with no public explanation, "model-as-a-service" starts to look like a liability — and that is a structural point every open-weights competitor can make without saying a word.

What This Signals About the New AI Export Regime

The Fable 5 shutdown is not an isolated incident. It is a preview of how the US government intends to regulate frontier AI as a strategic technology. The Commerce Department's ability to issue a directive that effectively pulls two of the world's most capable AI models from global circulation — without a court order, without a congressional process, and with no public explanation — is a demonstration of administrative power that no one in the AI industry had fully priced in before last week. Anthropic's legal structure, its San Francisco headquarters, its federal funding relationships: all of these give the US government levers that no Chinese AI company's customer has to worry about when working with a domestically hosted model. That asymmetry is now explicit, and every non-US Anthropic customer is recalculating their vendor risk accordingly.

The Relaunch Question

The models will almost certainly come back. The commercial and diplomatic costs of leaving them offline are too high for either side to absorb indefinitely — Anthropic loses revenue, the administration loses a relationship with its most strategically important AI company, and Korea's tech sector starts treating US AI vendors as unreliable. But the terms of the relaunch will matter more than the relaunch itself. If Anthropic is required to implement nationality filtering before restoration, it sets a precedent for every frontier AI company operating internationally. If the jailbreak issue is resolved without policy change, the White House looks like it blinked. And if the models return exactly as they were, the question becomes what the last seven days were actually about — and whether it will happen again.

Sources: Anthropic: Fable 5 launch · Anthropic statement on the directive · The Decoder (via WIRED) · The Register · Constellation Research · David Sacks on X · Tech Times · Kalshi