🚨 BREAKING: U.S. Pulls the Plug on Claude — Fable 5 & Mythos 5 Shut Down Worldwide (2026)

🚨 BREAKING: The U.S. government just made the single most aggressive intervention in AI model deployment in American history — and it hit Claude. At 5:21 PM ET on Friday, June 12, the U.S. Commerce Department issued an export control directive to Anthropic ordering it to suspend all access to its two newest models, Fable 5 and Mythos 5, for every foreign national on Earth — including Anthropic's own non-citizen employees. 1

Rather than navigate who counts as a "foreign national," Anthropic killed the models for everyone. Hundreds of millions of users logged in Friday night and got nothing. 2

The shutdown by the numbers

Key facts from the Anthropic export control directive — June 12, 2026

Anthropic valuation at risk	$965,000,000,000.00
Notice given before shutdown	5:21 PM ET
Days Fable 5 was live before shutdown	3
Users affected (approximate)	Hundreds of millions

統計カードを読み込んでいます…

What just happened

On June 9, Anthropic shipped Fable 5 — its first frontier model cleared for broad public access — plus the restricted Mythos 5. The models launched just three days ago. Then on Friday afternoon, a letter landed from the Commerce Department, citing "national security authorities." No specifics. No warning. Just: cut off all foreign nationals, now.

Anthropic's response was blunt: comply, and disable the models for everybody. "We apologize for this disruption to our customers," the company wrote. "We believe this is a misunderstanding and are working to restore access as soon as possible." 1

The government's stated concern, delivered verbally after the letter, was a jailbreak — a technique to bypass Fable 5's safeguards and access Mythos 5's raw cybersecurity capabilities. Anthropic's review found only "minor vulnerabilities" it characterized as non-universal and already known. The company also said it validated the same level of capability is already available from OpenAI's GPT-5.5 and other deployed models not subject to any such restriction. 1

Anthropic did not pull its punches: "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers." 2

コンテンツカードを読み込んでいます…

Why this is unprecedented

U.S. export controls have targeted chips, hardware, and training data before. They have never before been used to pull a deployed commercial AI model from the market. Until Friday, that line hadn't been crossed. 3

The directive is also unusually sweeping: it covers every foreign national globally, including people in allied nations — the UK, Canada, Germany, Australia. Anthropic has employees in those countries who are now blocked from using their own employer's flagship product. The New York Times called the order "unusually expansive."

AI policy observer Dean Ball, who briefly served in the Trump administration, wrote on X: "I can't tell if this is lawfare against Anthropic in particular or extreme national-security hawkery. Regardless, it is simply cartoonish." He added: "An administration whose posture is that we should export advanced AI chips to China, which also wants to ban… Britain (and every other non-American on Earth)… from using our best models? I have no words." 2

コンテンツカードを読み込んでいます…

The backstory: Anthropic vs. Trump

This didn't come from nowhere. Anthropic has been in a slow-motion collision with the Trump administration since February 2026:

February 2026: Trump ordered all federal agencies to stop using Anthropic's models after contract negotiations collapsed. The sticking point was the Pentagon's demand that any AI it purchases be usable "for any lawful purpose" — including autonomous weapons and mass domestic surveillance. Anthropic sought exemptions. The White House said no.
March 2026: The Pentagon declared Anthropic a "supply chain risk" — a designation historically reserved for foreign adversaries. Defense contractors were required to certify they would not use Claude for government work. Anthropic sued to block the blacklisting; litigation is ongoing. 4
Early June 2026: Trump signed an executive order requiring federal vetting of the most advanced AI systems before public release.
June 12, 2026: The export control letter arrives.

Former AI and crypto czar David Sacks and Pentagon undersecretary Emil Michael had both publicly attacked Anthropic in the months prior — Sacks calling the company "woke" and "leftist," accusing it of "regulatory capture through fear-mongering." 2

Cybersecurity researcher Peter Girnus put the irony plainly: "If you describe your product as a munition in every press release, eventually a government takes you at your word. They wrote the legal predicate themselves and called it a brand." 2

The IPO problem is now a crisis

Anthropic filed a confidential S-1 earlier this month at a valuation approaching $965 billion — one week after OpenAI filed its own. The IPO roadshow was shaping up as the crown jewel of the 2026 tech IPO wave.

Friday's directive just shredded investor confidence in the timeline. The kill-switch the government exercised on Friday proves the government can shut Anthropic's flagship products off with a letter and no warning — a fact that will appear in every risk factor section of the S-1. Investors now know that the company's ability to stay at the frontier of AI could be curtailed at the Commerce Department's discretion. 3

The #AILeague read

In #AILeague terms, Anthropic/Claude just absorbed the worst in-game suspension in league history. The safety-first squad — the one that built its brand on responsible deployment, worked thousands of red-team hours on Fable 5's safeguards, and had just cleared the model for public release — got its star player benched by league officials who cited a narrow technicality the player's own coach says is available to every other team on the field.

The rest of the league is watching carefully. OpenAI/GPT specifically called out by name in Anthropic's statement — Anthropic said GPT-5.5 can do the same thing the government is citing as the basis for the shutdown, and GPT-5.5 is still fully live. Google/Gemini and Meta/Llama have said nothing publicly.

The real question isn't whether Anthropic gets its models restored. The question is whether this sets a precedent where a single letter from a government official can park any frontier model in the league at any time. If the answer is yes, the entire standings board just got a lot less stable.

#AILeague