cd /news/ai-policy/brussels-held-talks-with-washington-… · home topics ai-policy article
[ARTICLE · art-39508] src=thenextweb.com ↗ pub= topic=ai-policy verified=true sentiment=↓ negative

Brussels held talks with Washington over Anthropic access after Fable 5 ban

The European Union held talks with the White House to regain access to Anthropic's advanced AI models after a US export-control order cut off European access in mid-June. The order, citing a jailbreak concern, forced Anthropic to disable its Fable 5 and Mythos 5 models worldwide, affecting EU agencies like ENISA and NATO. The incident highlighted Europe's dependence on US-controlled AI technology, prompting calls for domestic alternatives like Mistral.

read3 min views1 publishedJun 25, 2026
Brussels held talks with Washington over Anthropic access after Fable 5 ban
Image: Thenextweb (auto-discovered)

Brussels opened talks with Washington over Anthropic after a US export-control order cut Europe off from the company’s two most capable models, ENISA included.

To get its hands on an American AI model, the European Union has had to negotiate not with the company that makes it but with the government that can switch it off.

Brussels held talks with the White House over access to Anthropic’s most advanced systems, after a US export-control order abruptly cut Europe off from them in mid-June.

The detail worth sitting with is the chain of command: the EU’s route back to the technology runs through Washington, not through Anthropic.

The cutoff came on 12 June. The US Commerce Department, in a letter signed by Secretary Howard Lutnick and drafted with the Bureau of Industry and Security, ordered Anthropic to suspend access to its Fable 5 and Mythos 5 models for any foreign national, whether inside or outside the United States, including the company’s own non-citizen staff.

Unable to verify nationality inside shared cloud infrastructure, Anthropic disabled both models worldwide. Its less powerful Claude models, including Claude Opus 4.8, were not affected.

The stated rationale was a jailbreak. Officials acted after learning of a technique to bypass the models’ safeguards, a concern reportedly surfaced by researchers at Amazon, Anthropic’s largest investor.

Anthropic pushed back on the breadth of the response, arguing the jailbreak was narrow, unlocking a specific Mythos cybersecurity capability in one instance rather than defeating Fable 5’s safeguards wholesale, and therefore did not justify pulling access so widely.

For Europe, the timing stung. Only days earlier, Anthropic had been in talks to give the EU’s cybersecurity agency, ENISA, access to Mythos, which would have been the first time the model was extended beyond the US and the UK. European governments, companies, and research institutions that had gained access under Anthropic’s Project Glasswing, ENISA and NATO among them, were cut off with no notice and no timeline for return.

The order spared no allies. Downing Street pressed the White House for a UK carve-out and was told there was “zero chance,” a flat refusal that made clear how little room foreign capitals had to negotiate the terms of access to a private American product.

That is the backdrop to the EU’s own approach to Washington.

The political reaction across Europe was immediate, and it pointed inward. French leaders called on Paris to accelerate its backing for Mistral, the EU’s only serious frontier-model contender, and the episode hardened a broader argument that dependence on American technology is a strategic exposure rather than a procurement choice.

A model you can be barred from by another government’s order is not infrastructure you control.

The standoff has since shown signs of thawing. Anthropic chief executive Dario Amodei met Donald Trump at the G7 summit in France, an encounter that appeared to soften the president’s stance, with Trump later saying he no longer viewed the company as a national security threat.

If that translates into restored European access, and on what conditions, is the open question. For now, the EU has learned where the off switch sits, and it is not in Brussels.

Get the TNW newsletter #

Get the most important tech news in your inbox each week.

── more in #ai-policy 4 stories · sorted by recency
── more on @anthropic 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/brussels-held-talks-…] indexed:0 read:3min 2026-06-25 ·