At 5:21 p.m. Eastern on June 12, Anthropic received a letter from the Commerce Department that gave the company no choice: disable Claude Fable 5 and Claude Mythos 5 for every customer, immediately, worldwide. The U.S. government, citing national security authorities, issued an export control directive to suspend all access to the models by any foreign national—whether inside or outside the United States, including Anthropic's own foreign employees—forcing the company to abruptly disable both models for all customers.
The letter did not provide specific details of its national security concern.
Anthropic indicated that the government's concerns relate to a possible method of bypassing, or "jailbreaking," safeguards intended to limit Fable 5's use for certain cybersecurity-related tasks, including identifying software vulnerabilities. The company disputes the severity of the finding. But the damage was done. Claude Fable 5, described by Anthropic as its most capable widely released model with state-of-the-art results on nearly all tested benchmarks, runs with always-on adaptive thinking, a 1-million-token context window, and 128K output tokens; Claude Mythos 5 shares the same capabilities without the safety classifiers and is offered only in limited availability through Project Glasswing. Both went dark within hours.
This was not a polite request. It was the first time the U.S. government has used export controls to shut down an American AI company's frontier models in real time—and it happened less than a week after Anthropic launched Fable 5 to the public.
What Does This Mean for the AI Arms Race?
The timing is brutal. Anthropic filed to go public on June 1 at a $965 billion valuation on roughly $47 billion in annualized revenue; the company passed OpenAI in revenue in April 2026.
As of June 2026, Claude Mythos Preview leads reasoning benchmarks with a score of 70.2, followed by Claude Fable 5 at 66.3. Anthropic was winning. Now its two most powerful models are offline, indefinitely.
Meanwhile, the rest of the frontier is moving fast. Four of the five largest venture rounds ever recorded were closed in Q1 2026, with frontier labs OpenAI ($122 billion), Anthropic ($30 billion), xAI ($20 billion) and Waymo ($16 billion) collectively raising $188 billion; the first quarter of 2026 was unlike any other for venture investment, driven by unprecedented spending on AI compute and frontier labs, with investors pouring $300 billion into 6,000 startups globally.
xAI released Grok 4.3 on Amazon Bedrock, bringing general availability to AWS users with a 1-million-token context window, configurable reasoning levels, and low-hallucination enterprise AI performance; Grok 4.3 achieves the lowest hallucination rate among frontier models and offers configurable reasoning efforts (none, low, medium, high).
Meta introduced Llama 4 Scout and Llama 4 Maverick, the first open-weight natively multimodal models with unprecedented context length support and the first built using a mixture-of-experts architecture, and previewed Llama 4 Behemoth, one of the smartest LLMs in the world.
The export control directive doesn't just hurt Anthropic. It sends a signal to every lab: the U.S. government can and will pull the plug if it believes a model poses a national security risk, regardless of commercial consequences or IPO timing.
Can Europe Fill the Vacuum?
Mistral AI is in talks to raise about 3 billion euros (about $3.5 billion) at a valuation of about 20 billion euros (about $23.1 billion), a figure that could go higher depending on investor demand, according to Bloomberg; the French AI startup's discussions with investors are at an early stage, and the terms could change.
Hot off its first corporate showcase in Paris, Mistral, the crown jewel of France's AI industry, is making a play for American banks, businesses and tech bros; founded by a trio of machine learning engineers who worked at Facebook and Google, the company has surged in recent years to become the main European contender in the race to build generative AI systems.
Mistral CEO Arthur Mensch reacted on social media saying the Paris-based company "got put in the spotlight in the last few days." The subtext is clear: when the U.S. government shuts down American models, European alternatives suddenly look a lot more attractive—especially to customers outside the United States who don't want their AI infrastructure subject to Washington's veto.
But Mistral faces its own constraints. The French startup has delivered leading AI models under compute constraints; backing from the French government, Nvidia, and ASML will give it access to much more compute. The question is whether Europe can move fast enough to capitalize on Anthropic's forced retreat.
