Bitter: The Sovereignty-Cooperation Paradox Shaping the Digital Future of the G7

Analytical Studies > Analysis on the 2026 Evian Summit
G7 Summits > 2026 Evian Summit > Road to the 2026 Evian Summit

The Sovereignty-Cooperation Paradox Shaping the Digital Future of the G7

Raphaël Bitter, G7 Research Group
July 15, 2026

Ever since the 2009 G7 L'Aquila Summit, commitments pertaining to the digital realm have occupied an increasingly prominent place on the G7 agenda. Their share of total commitments has grown at each subsequent summit, reaching a robust 47% of all commitments made in the leaders’ declarations in Kananaskis in June 2025.

The 2026 summit in Evian on June 15–17 broke this trend. It saw a decrease in the proportion of digital-related commitments, largely due to the high number of commitments adopted overall and the prominence of ongoing geopolitical crises and economic challenges in this year’s communiqués. However, what the Evian Summit lacked in the quantity of digital commitments, it more than compensated for in their substance.

For the first time, founders and chief executives of several artificial intelligence companies including OpenAI, Anthropic and Google DeepMind attended a working session with G7 leaders on the summit’s final day to discuss access to frontier models and cooperation on matters where AI poses potential threats of cyberterrorism and national security. The presence of tech leaders at the summit is a testament to the substantive effort made by the G7 to tackle these issues. Soon after the summit, reports of initiatives created at this working session began to emerge as host French president Emmanuel Macron stated that he expected progress in the coming weeks on broadening access to leading US AI models. Indeed, one such reported initiative was the implementation of a US-led coalition on AI through a “trusted partner” program.

However, this spirit of cooperation is only one dimension of the broader dynamic that governed the discussions on digital technologies at the Evian Summit. Indeed, the attendance list for this working session revealed an underlying willingness for technological sovereignty on the part of the non-US G7 members. In addition to the heads of the so-called “big three” AI hyperlabs previously mentioned, the chief executives of leading AI companies of every member in the G7 participated, plus one from India. This deliberate inclusion reflected a growing consensus that AI capabilities are deeply intertwined with the future development of all G7 economies, and that dependence on the US is a liability none wish to be subject to.

This consensus and underlying rhetoric of technological sovereignty, already latent within G7 digital discussions, was accelerated by the US blockage of certain frontier models just days prior to the summit, thereby bringing concerns about technological sovereignty to the forefront of G7 discussions. On June 12, the US government had ordered Anthropic to suspend two of its frontier models, citing concerns of “national security.”

Anthropic complied, shutting down these models on a worldwide scale, without advance notice. The sudden suspension of these frontier AI models at US discretion highlights the current dependence of G7 members on the US with regards to AI infrastructure, and raises several questions about navigating the nexus between sovereignty and cooperation.

Perhaps the biggest of these questions is what comes next. It is apparent from the G7 summit in Evian that G7 members do not appear poised to accept a tradeoff in this apparent dilemma, and fully intend to pursue sovereignty and cooperation simultaneously. On the cooperation front, in Evian, leaders unanimously agreed to the “sharing of data, impartial evaluations and common standards in assessments methodologies of artificial intelligence models and algorithmic systems.” Furthermore, they agreed to rely on the “G7 cyber expert group to, as appropriate, enhance information sharing and identify best practices, in light of the recent developments regarding frontier artificial intelligence models [and] encourage further dialogue between cybersecurity agencies and relevant institutions in existing G7 groups.” Simultaneously, G7 members have already started implementing initiatives to increase technological independence from the US. Programs such as the EU’s InvestAI, France’s PROQCIMA and the UK’s ProQure, and Italy signing the Volta Declaration, are a testament to this, with hundreds of billions of US dollars being directed to this end.

Whether these two tracks can successfully be pursued simultaneously in the long term remains to be seen. But the G7 has positioned itself as the primary forum for coordination between some of the world’s most advanced economies and the world’s most powerful AI companies, forming a framework that excludes China altogether, making the framework’s claim to universal legitimacy the one question that Evian left unanswered.

[back to top]

—	This Information System is provided by the University of Toronto Libraries and the G7 Research Group at the University of Toronto.
	Please send comments to: g7@utoronto.ca This page was last updated July 15, 2026.