Lagarde urges Europe to build sovereign AI and data-centre capacity

A warning about strategic dependence

European Central Bank President Christine Lagarde has urged Europe to develop domestic artificial-intelligence models and substantially expand its data-centre capacity. Speaking in Vienna on September 14, she framed the issue as economic sovereignty: Europe risks either delaying adoption because it cannot adequately control sensitive data or embracing foreign systems so quickly that critical parts of its economy become dependent on suppliers based elsewhere.

Lagarde argued that a future interruption or repricing of foreign AI access could affect many sectors simultaneously. Automated systems are expected to influence border screening, tax administration, transport, healthcare and payments. Dependence at that scale, she said, could give an external trade partner unusual leverage during disputes over tariffs, digital taxation or other economic policies.

Europe's infrastructure gap

Her remarks highlighted the concentration of advanced AI development and computing capacity in the United States and China. The Guardian reported that Lagarde contrasted dozens of notable models produced in those two countries with one each in France and Britain, while saying the United States hosts roughly three-quarters of global AI computing capacity and Europe about five percent. These figures describe a structural gap rather than an immediate supply cutoff.

The policy direction is already visible in European Commission proposals. The official Cloud and AI Development Act proposal seeks to strengthen the European cloud and AI ecosystem, accelerate data-centre construction and reduce exposure to a small group of non-European providers. It aims to at least triple EU data-centre capacity within five to seven years while encouraging resource efficiency and sovereign cloud services for critical public-sector uses.

Growth opportunity and implementation risks

Lagarde paired the security argument with a productivity case, estimating that rapid AI adoption could raise European productivity materially over a decade. Yet expanding computing infrastructure creates difficult trade-offs. Data centres require large amounts of electricity, water, grid capacity and capital, while permitting delays and uneven infrastructure could concentrate investment in a small number of established hubs.

The next test is whether EU institutions and national governments convert the strategy into approved projects, financing and faster grid connections without weakening environmental safeguards. Policymakers must also determine which services require European-controlled infrastructure and which can safely continue using foreign providers. Lagarde's intervention does not itself create new law, but it adds the ECB president's economic-security authority to an existing Commission push for greater technological autonomy.