Europe stands to forfeit control of its AI destiny by 2031 unless leaders move swiftly on compute investment, talent development, and workforce protections. That warning anchors a new scenario report from a coalition of European researchers who modeled the continent’s trajectory against current policy and spending patterns.
The stakes have sharpened considerably as compute capacity increasingly dictates which nations get priority access to frontier AI models. Without intervention, Europe’s position barely improves from its current 5% share of global AI compute, while the United States commands roughly 70%.
Daan Juijn, Stan van Baarsen, Judith Dada, and five collaborators authored the report, titled Europe 2031. The format mirrors AI 2027, the April 2025 analysis from the AI Futures Project that mapped worldwide AI capability growth. This new work narrows the lens to Europe alone, testing what happens if investment and policy remain unchanged.
Today, Nordic countries house most European compute. The authors expect that distribution to shift by 2031, with France (37%), Norway (9%), and Germany (9%) emerging as the continent’s largest holders.
The contributors worked independently, with Arq Foundation covering website and translation costs. No AI company funding supported the effort. Juijn directs research at Arq Foundation, van Baarsen co-wrote the Dutch National AI Plan, Dada advises Germany’s government on AI transformation, and co-author Michiel Bakker works as an MIT assistant professor and Google DeepMind researcher.
Among five recommendations, the report urges a Danish-style flexicurity model. Employers would deploy AI aggressively while workers receive retraining and income support instead of job protections. The authors argue that shielding roles from automation only delays disruption, leaving workers without safety nets when faster-moving competitors prevail.
Governance gaps compound the problem. Many European civil servants lack firsthand experience with frontier AI systems due to data-protection restrictions. Consequently, regulators without technical fluency must oversee technology they cannot fully evaluate. Companies face a parallel risk when executives restrict AI access without building workforce AI literacy, then make decisions about jobs and skills without understanding what the tools actually do.
