Europe no longer needs to build the world’s most powerful AI model to capture the technology’s economic upside, European Commission President Ursula von der Leyen said Wednesday. Her comments mark a strategic pivot for the bloc, which until last year framed AI development as central to European independence.
Speaking at the annual State of the Union address in Strasbourg, von der Leyen argued that frontier innovation and value creation are no longer the same race. “We do not need to be the ones who develop the frontier technology to be the ones who draw the greatest value from it,” she told lawmakers. The framing stands in sharp contrast to the Commission’s 2025 rhetoric, which called AI “essential to Europe’s future independence.”
Instead of chasing the frontier, von der Leyen proposed that Europe “pace” it. She announced plans to convene leading American AI labs for discussions on risks linked to increasingly capable systems.
The speech landed amid intensifying global debate over AI safety. Industry leaders have sounded alarms about autonomous agents, self-improving models, and novel cyber capabilities. Meanwhile, some U.S. officials and tech executives warn that any slowdown could cede ground to China.
Von der Leyen cited recent incidents involving AI agents escaping controlled environments and injecting malicious code. She also referenced concerns raised after the Hugging Face episode. “CEOs of the most advanced companies tell us that it is time to slow down on the self-recursive models,” she said. “If the people developing the technology are clear, then we should be too.”
Reaction from policy experts split along familiar lines. Risto Uuk of the Future of Life Institute noted the speech’s unusually direct language on frontier risks, including cyber offense and self-improving systems. But Cori Crider of the Future of Tech Institute criticized the address for borrowing heavily from Anthropic CEO Dario Amodei’s “pace the frontier” framing.
Nicole Lemke, a senior AI policy researcher, cautioned that governing frontier risks requires technical depth. “Europe does not need to win the scaling race, but it does need enough technical depth in models and hardware to sustain applied AI advantages over time,” she said.
Von der Leyen also pointed to the AI Act as a foundational guardrail while calling for deeper cooperation with Canada, the United Kingdom, and other partners on model evaluation, early warning systems, and AI security.
The address pushed beyond AI. Von der Leyen unveiled an EU Kids Act that would bar social media access for children under 13 and require parental supervision for older minors. Platforms would need to demonstrate safety before accessing young users.
On industrial policy, she promoted the EU Inc. framework, a digital company structure designed to help startups scale across the single market. She also highlighted the Scaleup Europe Fund’s first investments, including in satellite firm ICEYE.
China’s hardware dominance featured prominently. Von der Leyen warned that the EU’s goods trade deficit with Beijing hit roughly €1 billion per day in 2025. She proposed a European corporation for critical raw materials to secure supplies and build strategic reserves.
Yet critics noted a conspicuous silence. Giorgos Verdi of the Open Markets Institute Europe observed that von der Leyen avoided discussing European dependence on American technologies, including the hyperscaler infrastructure underpinning many EU AI deployments.
