Christine Lagarde says Europe needs its own AI before trade partners use it as a weapon

Key points

• ECB president Christine Lagarde warned that Europe's dependence on foreign AI technology gives trade partners direct power over the continent.
• Lagarde called for European-built AI models, the software behind chatbots and other AI tools, capable of handling most everyday tasks.
• She argued that domestic data centres, the large warehouses of computers that run AI services, are essential to make that independence real.
• Europe currently relies heavily on AI infrastructure controlled by US and Chinese companies.

Christine Lagarde does not do vague. When the president of the European Central Bank speaks about AI dependency as a geopolitical liability, the conversation has moved well past the tech pages.

Speaking publicly, Lagarde said Europe needs AI models, the underlying technology that powers tools like chatbots and image generators, that are "good enough" to handle the majority of tasks its businesses and governments need done. Those models must run from data centres on European soil, not from servers owned by American or Chinese firms.

Her core argument is simple: if a trade partner controls the AI your economy depends on, they can threaten to switch it off. "The threat of being cut off loses its force," she said, if Europe builds its own alternative.

Why does this matter to ordinary people?

Because the services quietly running in the background of daily life, from automated customs checks to hospital record systems to the chatbots answering your tax questions, increasingly depend on AI software built and hosted abroad.

If that infrastructure sits in data centres in Virginia or Shanghai, a government in Washington or Beijing holds a lever that has nothing to do with tariffs or sanctions. It is a newer kind of pressure, and trade negotiators are starting to price it in.

First reported by The Guardian, Lagarde's comments arrive as the European Union is already spending heavily on the AI Factories programme, a plan to build publicly accessible supercomputers for AI research across the continent. Her remarks suggest that effort needs to go further, reaching into commercial AI development and the physical data centre capacity to run it at scale. That pressure on Europe to act more urgently is something we traced back to TechBBQ in Copenhagen on 31 August, where the same question about foreign dependency ran beneath every panel.

Europe has world-class AI researchers. What it has lacked is the concentrated private investment that built OpenAI and Google DeepMind. Lagarde is essentially arguing that closing that gap is no longer just an economic question. It is a matter of sovereignty.

The "good enough" framing is deliberate. She is not calling for Europe to beat the US at frontier AI, the most advanced models that push performance records. The goal is models that are reliable, local and fit for the tasks European institutions actually need, a more achievable target and a more politically honest one.

What happens next?

No binding proposal sits behind Lagarde's remarks yet. They are a signal of direction, not a law or a regulation. Watch the European Commission's AI investment roadmap, due later this year, for whether political will follows the rhetoric.

Common questions

Does Europe already have its own AI models?

Yes, but they are smaller in scale and funding than US or Chinese rivals. French company Mistral AI and a handful of academic labs have released competitive open models, but no European AI service matches the reach of OpenAI or Google at this point.

What is a data centre, and why does its location matter?

A data centre is a large building filled with powerful computers that store and process information. If your AI service runs on computers in another country, that country's government can legally compel the company to restrict or cut access, which is why Lagarde says Europe needs its own.