Lagarde sets Europe's AI choice as forgone growth or dependence on US and Chinese tools

Two sets of documents govern a European company's use of AI, and they are written in different places. The capability roadmap comes from the vendors, and the Guardian's reading is that their pace cannot be changed without Washington or Beijing, neither of which has agreed to slow anything ^[4]. The compliance obligations come from Brussels. Margrethe Vestager, the EU's former competition commissioner, said Europe does have a voice in the safety debate through the EU AI Act. The Act sets requirements for developers of "high-risk" systems to report safety incidents and to address risks such as a cutting-edge model evading control ^[10].

The absence at the top of the safety debate follows from scale. The Guardian traces it to the continent's longstanding inability to produce a Silicon Valley-scale titan like Apple, Google or OpenAI, and to the lack of an Anthropic or Nvidia-sized frontrunner, while noting that Europe is a powerhouse in regulation, much to the White House's chagrin ^[2]. All five companies named as the benchmark for scale are American ^[19].

Both remedies on offer take years. Lagarde said on Monday that Europe must develop its own AI technology and build more datacentres to nullify the threat of being cut off by the US or China. If it invests, she said, "the threat of being cut off loses its force" ^[5][6]. Vestager co-authored a report this week, A Transformative AI Strategy for Europe, that asks for the same datacentres and for resilience enough to withstand "AI crises" like a powerful model slipping out of control ^[7][9]. It says European citizens "are already bearing the economic and social consequences of decisions taken elsewhere" ^[8]. The Guardian's account does not put a cost or a date on the build. The datacentres answer a dependency measured in decades. The contract a European buyer signs this quarter is governed by the vendor's terms and the AI Act as it stands.

The skeptic's case is made in the same piece. Frederike Kaltheuner, an adviser to the AI Now Institute, says the narrative that Europe is grievously behind depends on a view of artificial intelligence aggressively promoted by American tech bosses ^[16]. Because of those doomsday warnings, she said, "we are exclusively talking about AI in a very narrow, specific kind of way" ^[15]. She also says high costs and fears over what might happen to their data are keeping many companies from adopting frontier models. If that continues, Europe may not be so terribly positioned, she says, because the AI Act carries pragmatic measures for how consumers meet AI day to day ^[17][18]. She still describes Europe as dependent on US technology in ways that will be hard to undo, including search engines, chips and datacentres ^[14].

The two views point at different budgets. If frontier access is the binding constraint, Lagarde's datacentres are the answer; if cost and data risk are what keep buyers off frontier models, the AI Act's consumer provisions matter more than the compute ^[5][17].

Vestager wants the existential question handled outside the two-power frame. She said Europe must "build its own AI to strengthen the ecosystem, retain talent, and inspire innovation" ^[11]. A threat to humanity, she said, needs a global solution that does not hand power to the US and China, and that takes a different kind of supranational body ^[13]. She pointed at the UN general assembly ^[13]. Her own words on the rivalry: "It is vital to move beyond the US-China rivalry narrative" ^[12].