What happened
RAND published this report on 7 July 2026, providing a scenario-based framework for assessing the conditional value of US-EU AI partnerships amid intensifying US-China competition. The analysis identifies four archetypal scenarios for how AI development could unfold, defined by two variables — the locus of AI value creation and the degree of advanced model openness — and evaluates which forms of transatlantic cooperation generate returns under each. Key findings include that evaluation/testing/standards interoperability and lithography resilience cooperation should be pursued 'now' regardless of scenario, while the report also argues the EU AI Act's compliance burden is driven more by GDPR and sectoral rules than the Act itself, and that US private capital reshaping EU AI infrastructure 'may foreclose broader access outcomes' the alliance should secure. Author Marzia Giambertoni concludes existing transatlantic institutions have not produced a genuine 'strategic AI partnership.'
Why it matters
Provides executives and policy leads a structured way to stress-test AI-related transatlantic dependencies (compute, model access, standards) against multiple plausible futures, informing decisions on where to hedge versus commit to US or EU-aligned AI infrastructure and partnerships.
Action needed
Map current AI vendor/infrastructure dependencies against the report's four scenarios to identify exposure to US-EU policy divergence, and flag the evaluation/standards and hardware-resilience cooperation areas as near-term engagement priorities.