G7 AI Sovereignty Push: 9 CEOs Convene as US Model Ban Exposes Legal Risks

For legal professionals, the G7 AI gathering is a wake‑up call: US export controls can abruptly cut off foreign access to critical AI infrastructure, creating unprecedented compliance and contractual challenges. The European Commission’s sovereignty package seeks to address this, but questions remain about enforcement, liability, and international trade law.

The June 17, 2026, working lunch on the sidelines of the G7 summit in France crystallized a tectonic shift in global AI governance, as Europe accelerates its push for technological sovereignty in the face of recent U.S. export controls that abruptly barred non‑Americans from accessing some of the world’s most advanced AI models. The gathering brought together nine chief executives from a cross‑section of leading AI firms—OpenAI’s Sam Altman, Google DeepMind’s Demis Hassabis, Anthropic’s Dario Amodei, and leaders from Canada’s Cohere, France’s Mistral, Germany’s Black Forest Labs, Italy’s Domyn, Japan’s Sakana AI, and the UK’s Synthesia—signaling both the breadth of the industry’s global footprint and the deep anxieties over the concentration of AI power in a handful of American companies.

“

On one hand, it could fuel a wave of investment in European AI startups like Mistral and Black Forest Labs, whose valuations are already buoyed by the promise of state‑backed procurement.

The immediate catalyst for the sovereignty debate was Anthropic’s suspension of its most advanced models, Fable 5 and Mythos 5, following an order from the Trump administration that cited national security concerns. The order not only barred non‑Americans outside the United States but also imposed restrictions on non‑American citizens working inside the country, forcing Anthropic to halt access globally. The episode vividly demonstrated the geopolitical leverage embedded in proprietary AI infrastructure: a single executive order can, overnight, render entire nations dependent on that infrastructure vulnerable. As Zach Meyers, director of research at Brussels‑based think tank CERRE, put it, “Europe, Canada or other countries can be put in an extremely vulnerable position if they get cut off from advanced AI models.”

The incident placed a sharp edge on the European Commission’s technology sovereignty package, unveiled earlier in June 2026, which aims to fortify domestic AI capabilities through funding, data‑sharing initiatives, and regulatory reforms. The package, coupled with Pope Leo XIV’s recent call for stronger AI oversight, reflects a multi‑layered response to the perceived risks of over‑reliance on foreign—primarily U.S.—AI technologies. The EU’s strategy is not merely about building competitive alternate models; it is also about creating a legal and operational framework that insulates its member states from unilateral actions by other governments. This includes considering mandatory fallback clauses in procurement contracts, mandatory data localization for sensitive sectors, and a possible expansion of the EU AI Act to include instruments that mitigate extraterritorial technology controls.