US lifts ban on Anthropic AI models in Europe

Anthropic AI's most advanced models, Claude Fable 5 and Mythos 5, are available for use again in Europe, the company announced in an X post. The United States Department of Commerce has officially reversed its decision to block European access to the AI models.

The restriction, which began in mid-June, sent shockwaves through the global technology sector and exposed the vulnerability of foreign organizations relying on American software infrastructure. While the immediate crisis has resolved, the three-week blackout has permanently altered how European enterprises view their digital supply chains. Policymakers and executives are now grappling with the reality that access to critical computing tools can be revoked instantly by foreign regulators.

The ban reversal

The decision to lift the export ban came directly from the United States Department of Commerce, led by Commerce Secretary Howard Lutnick. To secure the reversal, Anthropic agreed to three strict security and regulatory conditions mandated by American authorities.

First, the company must proactively detect and address potential security risks associated with its frontier models. Second, Anthropic must coordinate closely with the United States government on established safety protocols, standards, and release procedures for all current and future models. Finally, the developer must immediately report any identified malicious activity to federal officials.

These conditions transition Anthropic into a highly regulated exporter, transforming how the company deploys its technology globally. While European users can now access the systems without an export license, the agreement establishes a precedent of direct state oversight over commercial artificial intelligence deployments.

The impact on European research

The sudden loss of access to Claude Fable 5 and Mythos 5 severely disrupted European scientific and academic communities. Because these models excel at complex, iterative problem-solving, many institutions had integrated them into active research workflows.

During the blackout, French Research Minister Philippe Baptiste sounded the alarm at the VivaTech conference, warning that the sudden "kill switch" threatened progress in mathematics and the advanced sciences. The incident highlighted a critical weakness: European laboratories lack equivalent domestic tools. Without access to these frontier systems, European researchers faced immediate competitive disadvantages compared to their American peers. The disruption forced universities and private laboratories to reassess their reliance on foreign-controlled software. Even though access has been restored, the academic community remains deeply wary of future regulatory interventions that could halt long-term scientific projects without warning.

The sovereignty wake-up call

For years, European policymakers have debated digital autonomy, but the Anthropic ban turned a theoretical risk into an immediate economic threat. Industry analysts point out that relying entirely on American-controlled artificial intelligence infrastructure creates a structural vulnerability.

The brief ban demonstrated that the United States government can unilaterally cut off foreign access to essential business tools. This realization has undermined trust in the "trusted partner" framework that previously governed transatlantic data and technology sharing. European business leaders now recognize that commercial contracts offer little protection against national security directives issued in Washington. The incident has intensified calls for Europe to build its own sovereign computing stack to avoid being at the mercy of foreign policy shifts.