TL;DR
At the G7 summit in Evian, European leaders pressed Dario Amodei, Demis Hassabis and Sam Altman over access to frontier AI models after a reported U.S. export-control order cut off Anthropic’s leading systems worldwide. The account says Europe wants durable access, trusted-partner rights, domestic AI capacity and child-safety rules, while the ban itself remains controlled by Washington.
European leaders used a June 17 G7 working lunch in Evian-les-Bains to seek more reliable access to frontier AI models from Anthropic’s Dario Amodei, Google DeepMind’s Demis Hassabis and OpenAI’s Sam Altman, after a U.S. export-control directive reportedly forced Anthropic to cut off its Fable 5 and Mythos 5 systems worldwide five days earlier. The dispute matters because European businesses and public institutions are confronting the risk that core AI tools can be interrupted by a foreign government order.
The working lunch, hosted by French President Emmanuel Macron during the June 15-17 G7 summit, put the leaders of Anthropic, Google DeepMind and OpenAI at the same table as heads of state. According to the summit account, President Donald Trump attended with Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnick and Secretary of State Marco Rubio; European participants included Macron, German Chancellor Friedrich Merz, European Commission President Ursula von der Leyen and U.K. Prime Minister Keir Starmer.
The immediate trigger was a June 12 directive from the U.S. Commerce Department that ordered Anthropic to block Fable 5 and Mythos 5 access for any “foreign national,” according to the account. Because nationality cannot be reliably checked in real time at API scale, the account says Anthropic applied a worldwide shutdown, cutting off European users without warning or a grace period.
The joint G7 statement framed the session around safe, rapid and effective AI deployment. The practical question at the table was narrower: how much Europe can rely on U.S.-based frontier models when access may be changed by the U.S. executive branch.
Évian and the fallout: what Europe actually wants
For the first time, Amodei, Hassabis, and Altman sat with heads of state — five days after Washington switched Anthropic’s models off worldwide. Europe’s question: can you rely on models a foreign cabinet can shut down by decree?
The dilemma: what Europe wants from the three CEOs, the three can’t deliver — because they don’t hold the switch, Washington does. Macron’s platform is the right answer, but no fix for a decade-old infrastructure gap. The only answer that doesn’t depend on someone else’s goodwill: your own models, your own compute, open weights you can self-host.
Europe Tests AI Dependency
The episode turns AI sovereignty from a policy slogan into an operational risk for banks, manufacturers, public agencies and software companies that have embedded hosted models into daily work. If access to frontier systems can be removed by a national-security order abroad, procurement, resilience planning and compliance all change.
For the AI companies, the lunch exposed a limit on what their executives can promise. Amodei, Hassabis and Altman can offer standards work, access programs and safety commitments, but export controls sit with Washington, not the labs.
The GPT-4 Millionaire: Future of Business Featuring Microsoft 365 Copilot: How to Leverage AI Language Models to Grow Your Company and How AI-driven Language Models Will Revolutionize the Way We Work
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Macron’s AI Sovereignty Push
Macron placed the G7 AI lunch within a wider European effort to build more domestic capacity. The source material cites a €420 billion package, AI gigafactories and the CADA initiative as parts of that agenda, alongside demands for a say in where compute, chips and power are located.
The guest list sent the same message. Alongside U.S. executives, the lunch included European and allied AI companies such as Mistral, Synthesia, Black Forest Labs, Domyn and Sakana AI, plus other technology leaders including Salesforce’s Marc Benioff and Meta’s Alexandr Wang.
“Democracies should not give in to the temptation to splinter.”
— Dario Amodei, Anthropic CEO
Hermes Agent: The Production-Grade AI Engine: Cut API Costs, Add Model Fallbacks, and Secure Your LLM Layer with LiteLLM
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Guarantees Remain Out of Reach
It is not yet clear whether any formal access guarantee can be created for European users, or whether Washington would accept binding limits on future model restrictions. The source material says the U.S. ban remained in place after the summit, with no reversal announced.
Several details also remain unresolved: how trusted partners would be defined, whether the plan would apply to companies, governments or both, and how child and youth safety rules would be aligned across countries.
Secure AI Model Deployment: A Comprehensive Guide to Safely Delivering Machine Learning Systems in Production Environments
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
September Talks Test Access Plan
Leaders are expected to reconvene in September for another Western-democracy meeting on AI, according to the source material. A separate platform is expected within one month, aimed at trusted-partner access, cyber defense against China-linked threats and shared child-safety principles.
The test will be whether those talks produce operational protections for European users, not just aligned statements. Until then, European governments and companies are likely to keep investing in domestic models, self-hosted open-weight systems and local compute as a hedge against another sudden cutoff.
Zero Trust Architecture Guidance: Rethinking Enterprise Security from the Inside Out (Agentic AI Enterprise Security)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What happened at the G7 AI lunch in Evian?
European leaders met with Dario Amodei, Demis Hassabis and Sam Altman on June 17, 2026, during the G7 summit in Evian-les-Bains. The meeting focused on AI access, safety standards and the political risk created by a reported U.S. order affecting Anthropic’s most capable models.
Why did Anthropic’s reported shutdown matter for Europe?
The account says European companies and public institutions lost access without warning after the U.S. Commerce Department directive. That made dependency on foreign-hosted AI systems a practical business and public-sector risk, not only a policy debate.
What did European leaders seek from the AI chiefs?
They sought reliable access to frontier models, trusted-partner rights for non-U.S. allies, protections against another sudden cutoff, a voice in AI infrastructure decisions, stronger domestic AI capacity and common child-safety principles.
Can the AI chiefs guarantee future access?
Not by themselves. The companies can shape product access, testing standards and partner programs, but U.S. export controls are government decisions. That is the central tension left by the Evian meeting.
What is expected next?
Western leaders are expected to meet again in September, with a platform due within one month. The main question is whether those efforts produce enforceable access protections or mainly shared political language.
Source: Thorsten Meyer AI
