Europe pushes for AI ‘sovereignty’ as it warns against depending on foreign models

European officials have increasingly framed reliance on foreign AI infrastructure and foundation models as an economic-security risk rather than a purely commercial preference, driving a broader push toward what’s being described as European AI “sovereignty” — independent compute capacity and homegrown models less dependent on US or Chinese providers.

The push sits alongside the EU’s continued enforcement of its AI Act, creating a two-track European AI strategy: regulate the use of AI systems within the bloc through the Act’s risk-tiered rules, while simultaneously trying to reduce dependence on the non-European companies that currently supply most of the underlying models and compute those rules govern.

Whether that sovereignty push translates into genuinely competitive European alternatives to US and Chinese frontier models remains an open question — building foundation models at the frontier requires enormous, sustained compute investment, and Europe has historically struggled to match the scale of capital that American AI labs have been able to raise. But the political appetite for trying appears to be growing, not shrinking, as AI becomes more central to economic competitiveness.