America just pulled Claude Fable 5 for all foreigners. Has the hybrid AI war begun?
On 12 June 2026 the US government ordered Anthropic to disable Claude Fable 5 and Mythos 5 for all foreign nationals. What does it mean for Europe and what can you do about it?
News analysis
Claude + Mistral
On Friday evening, 12 June 2026, Anthropic received an order from the US government: block Claude Fable 5 and Mythos 5 for all foreign nationals, wherever they are. Anthropic chose to switch them off entirely, for everyone. Three days after the grand launch, the best public model in the world is unavailable to the rest of the planet. What was written about as sci-fi only last week is here.
What's new
For the first time America has banned a specific AI model for foreigners in practice. Anthropic received an order from the Commerce Department to suspend access to Fable 5 and Mythos 5 for every foreign national, including people inside the US and its own staff without American citizenship. Because partial compliance would have meant cutting off a huge slice of users, Anthropic switched the models off for everyone at once.
The reason is security. The government says another company showed it could bypass the Fable 5 safeguards and reach the offensive cyber capabilities of the Mythos class, the ability to find and exploit weak spots in other systems. Anthropic argues it was a narrow jailbreak, but the decision had already been made. Other Claude models, such as Opus, keep running.
That turned theory into precedent. A frontier model behaves like military technology: first it ships to the world, then the government pulls it by citizenship with a single order. And that exact lever now hangs over anyone who built their operation on top-tier American AI.
What you'll like most about it
That it is not the end of the world, just a wake-up call. You lost one specific model, not AI itself. The other Claude models keep running and European and local alternatives are ready.
And it is a wake-up call at the right moment. Three days earlier the European Commission unveiled a technological sovereignty package meant to reduce dependence on American and Chinese technology. Suddenly it has a very concrete reason.
Who it's for
It hit hardest the companies that built their operation on a top-tier American model. Whoever was building on Fable 5 yesterday is hunting for a replacement today.
Company on a top model
- Product
- Operations
- Continuity
Creator and marketer
- Copywriting
- Visuals
- Campaigns
Team with sensitive data
- Data residency
- Compliance
- Security
How to use it in practice
Start by finding where a similar order could hit you. Go through your tools and flag the ones that rely solely on a single American model. That is your risk zone, as Fable 5 just showed.
01 · Map
02 · Diversify
03 · Keep a local backup
A real-world example
A Prague marketing agency ran its client output on Claude Fable 5. On Friday evening the model vanished without warning. Because it had a backup ready, on Monday morning it moved routine work to Mistral, sensitive documents to a local model on its own server and finished the rest on Opus. The clients noticed nothing. A neighboring agency that ran only on Fable 5 spent the whole week dealing with the outage.
Tools worth a look
Fable 5 just showed why your own hardware and a local LLM pay off. A model running on your own machine cannot be switched off by any government, repriced overnight, or blocked by passport. That is sovereignty you arrange yourself.
Local models
Qwen, Llama
Best for
Anyone who wants AI fully under control, independent of an outside company.
Mistral Large 3
Mistral AI
Best for
Companies and creators who want a European model and EU data.
European cloud
OVHcloud, Scaleway
Best for
Companies that need both operations and data physically in the EU.
Summary
This is no longer "what if". It happened. The best public model in the world was switched off for everyone outside the US three days after launch. Access to top-tier AI is now visibly decided by geopolitics, not by you.
The good news: Europe is finally pushing its own plan, and Mistral and local models show you have a choice. And your insurance is simple. Do not build everything on one provider, keep a European and a local backup on hand, and know where your data lives. Whether we depend on America again, or finally build something of our own, remains to be seen. What you can influence right now is your own dependence.
Sources
- Anthropic: Statement on the US government directive to suspend access to Fable 5 and Mythos 5
- Al Jazeera: US orders Anthropic to disable AI models for all foreign nationals (13 June 2026)
- Fortune: Anthropic disables Fable and Mythos following US export ban (13 June 2026)
- The New Stack: Federal government orders Anthropic to pull Fable 5 and Mythos 5, three days after launch
- Euronews: Europe launches a tech push to break US and China dependence (June 2026)
Frequently asked questions
What people often ask
Can I still use Claude Fable 5 from Europe?
No. Anthropic disabled both Fable 5 and Mythos 5 for all users after the government directive of 12 June 2026. The ban covers all foreign nationals, so a selective shutdown outside the US made no sense and Anthropic cut access entirely. Other Claude models, such as Opus, keep working normally. If you built on Fable 5, you have to move to another model. Whether and when Fable 5 returns, nobody has said yet.
Why did America ban Fable 5?
For security. The government says another company showed it could bypass the Fable 5 safeguards and reach the offensive cyber capabilities of the Mythos class, the ability to find and exploit weak spots in other systems. The Commerce Department judged it a national security risk and on Friday 12 June ordered access for foreign nationals suspended. Anthropic argues it was a narrow jailbreak, but the decision had already been made.
Does this also affect ChatGPT and Gemini?
Not yet. The order targeted Anthropic specifically and the Mythos-class models because of their cyber capabilities. ChatGPT, Gemini and the other Claude models were not hit. What matters, though, is that it revealed the mechanism: the government can pull a specific model by the user's citizenship with a single order. What happened to Fable 5 could next apply to another frontier model.
How do I prepare so the next ban does not hit me?
Do not run your whole operation on one American model. Map the tools you use daily and flag the ones that rely solely on a single provider. Then add a second model from another company as a backup, ideally Europe's Mistral. For sensitive data and maximum certainty prepare a local open weight model on your own hardware. Nobody can switch that off. When the next directive comes, you switch within a day instead of fighting an outage.
Keep going
Related articles
More guides from the same area, topics and tools.
China is betting 2 trillion yuan on AI without Nvidia. Will it build its own world?
China is preparing a five-year plan worth 2 trillion yuan (~$295bn) to build a national AI grid that runs 80% on domestic Huawei chips. What does it mean for the models you use?
OpenAI just shipped three new GPT-5.6 models. Why won't the US government let anyone near them yet?
OpenAI unveiled GPT-5.6 in three tiers, Sol, Terra and Luna. Because of a Trump administration intervention, only a handful of vetted companies can touch them so far.
SpaceX is already public, OpenAI and Anthropic are next in line. Will your AI usage drive their stock?
SpaceX just pulled off the largest IPO in history, with OpenAI and Anthropic queuing for trillion-dollar valuations. Will the share price come down to how much AI we actually use?