TL;DR
European leaders used a June 17 G7 AI working lunch in Evian to press AI chiefs on access, sovereignty and shutdown risk after a U.S. export-control directive hit Anthropic models, according to the supplied dispatch. The core issue is whether Europe can rely on frontier models controlled by firms subject to U.S. executive action.
G7 leaders met Anthropic’s Dario Amodei, Google DeepMind’s Demis Hassabis and OpenAI’s Sam Altman at an AI-focused working lunch in Evian-les-Bains on June 17, days after a U.S. Commerce Department directive forced Anthropic to block Fable 5 and Mythos 5 for foreign nationals and caused a worldwide shutdown, according to the supplied Thorsten Meyer AI dispatch.
The dispatch says French President Emmanuel Macron devoted the working lunch to artificial intelligence, with the official theme centered on safe, rapid and effective AI deployment. The guest list included U.S. President Donald Trump, senior U.S. officials, German Chancellor Friedrich Merz, European Commission President Ursula von der Leyen and U.K. Prime Minister Keir Starmer, alongside other technology leaders and allied AI labs.
Amodei, Hassabis and Altman arrived with overlapping proposals, according to the source material. Amodei pushed for a U.S.-led democratic coalition with structured frontier-model access for trusted partners, chip trade excluding China and joint defenses against AI risks. Hassabis backed a Western coalition. Altman proposed an international forum for testing standards and argued that companies alone should not set the rules.
The European side, as described in the dispatch, wanted more than broad pledges. Its demands included durable access to frontier models, protection against another sudden shutdown, a trusted-partner access scheme, more influence over compute and chip infrastructure, stronger European AI capacity and common child-safety rules. The dispatch says the Anthropic restrictions had not been reversed.
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 Foreign Model Dependence
The Evian meeting mattered because the Anthropic shutdown turned a policy debate into an operational risk for European companies and public institutions that had built foreign frontier models into daily work. If access can be removed quickly under another government’s order, procurement, public services and critical business systems face a new reliability problem.
The meeting also showed a power mismatch. The three AI executives can offer access frameworks, testing forums and voluntary cooperation, but the dispatch argues they cannot fully remove the shutdown risk because the switch sits with Washington. For businesses and investors, this is a policy and supply-chain risk marker, not financial, tax or legal advice.
Access Control Systems: Security, Identity Management and Trust Models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Shutdown Before Evian
According to the dispatch, the U.S. Commerce Department issued its directive on June 12, ordering Anthropic to block its most capable models for any foreign national. Because nationality cannot be reliably checked in real time at API scale, the source says Anthropic responded with a worldwide shutdown rather than a narrower block.
The source frames Macron’s AI platform and a reported EUR420 billion sovereignty package as part of Europe’s answer, alongside AI gigafactories and the CADA initiative. Those figures are attributed to the supplied material and describe policy positioning; the material does not show how much funding is enacted, budgeted or private capital.
“Temptation to splinter.”
— Dario Amodei, Anthropic CEO
Auditing Artificial Intelligence
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits Of Any Access Promise
Several details remain unclear from the supplied material. It does not provide the full legal text of the U.S. directive, its duration, appeal channels, the exact number of affected users or whether some services were restored later.
It is also unclear whether a trusted-partner scheme would be legally binding, who would qualify, and whether European governments could get enforceable protection against future U.S. export-control action. The scale, funding and timetable of Europe’s own model and compute buildout also remain open.
Train It. Tame It. Teach It.: Build Your Personal AI Team and Get Every Model to Work Your Way
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
September Talks Will Test Terms
The dispatch says Macron’s platform is expected within a month and that Western democratic leaders plan to reconvene in September. The next test will be whether governments can turn access promises into enforceable terms covering model availability, chip supply, cyber-defense cooperation and child-safety principles.
Until then, European institutions and companies are likely to keep reviewing contingency plans, including multi-model procurement, self-hosted systems and open-weight alternatives.
AI FOR CORPORATE GOVERNANCE & COMPLIANCE: Your Complete Implementation Guide to Transforming Governance from Compliance Cost Center to Strategic Advantage … & MANAGEMENT LIBRARY SERIES Book 17)
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?
According to the supplied dispatch, G7 leaders met major AI executives on June 17, 2026, to discuss AI deployment, safety and access after a U.S. directive triggered a worldwide Anthropic model shutdown.
Why did the Anthropic shutdown matter to Europe?
The shutdown showed that European users of frontier AI models can lose access because of foreign government action, even when the affected services are already built into business and public-sector workflows.
What did the AI CEOs propose?
The source says Amodei and Hassabis backed a democratic AI coalition, while Altman proposed an international forum for testing standards. Amodei also pushed structured access for trusted partners and chip controls excluding China.
What does Europe want now?
The dispatch lists reliable access, guarantees against another shutdown, a trusted-partner scheme, more say over compute infrastructure, stronger European AI capacity and child-safety rules.
Is the U.S. restriction still in place?
The supplied material says the ban had not been reversed in the immediate fallout from the summit. It does not give a later update on any legal change, exemption or service restoration.
Source: Thorsten Meyer AI
