Draghi Warns Europe Risks 'Catastrophic' Outcome Without AI Data Center Push

Former European Central Bank President Mario Draghi issued a stark warning that Europe must urgently build more artificial intelligence data centers to maintain its economic sovereignty and reverse a widening productivity gap with the United States. In an article for the Financial Times, Draghi argued that failing to secure independent AI infrastructure would have "catastrophic" effects on the region's economy.

The Widening Compute Gap

Draghi highlighted a significant and growing disparity in technological infrastructure between Europe and the U.S. The European Union currently hosts under 5% of the world’s AI compute capacity, compared with 75% held by the United States, he wrote.

This shortfall is creating a severe supply-demand imbalance. According to Draghi, the gap between demand and installed supply in Europe is already around 3 gigawatts (GW) and is projected to widen to 14GW by 2030. This technological lag contributes to a productivity gap with the U.S. that widened from $9 per hour in 2018 to an estimated $21 per hour in 2025.

Economic Sovereignty at Risk

Draghi stressed that dependency on foreign AI infrastructure poses a critical threat. "Being cut off from AI, once the economy runs on it, would be more like being cut off from the US financial system," he stated. The stakes are high, as the European Commission estimates Europe’s data economy could be worth over €800 billion, or more than 5% of GDP, by 2030.

Ad

According to scenarios from the ECB cited by Draghi, the rapid adoption of AI could add 0.3 to 0.4 percentage points annually to the region's total factor productivity growth, which has been near zero since 2022. Securing control over data and the infrastructure to process it is therefore essential for future economic growth.

A Call for Collective Investment

To overcome this challenge, Draghi pointed to both regulatory hurdles and viable solutions. Building a data center takes an average of 42 months in Germany, compared to just 24 months in the U.S., due to longer permitting and grid connection delays. However, he noted that cost is not the primary barrier, citing that an AI data center in Sweden costs only about 10% more than one in China.

Draghi proposed that European companies pool their resources to finance new data centers through large-scale, long-term contracts. He referenced an existing initiative where companies including ASML, Capgemini, and Amadeus committed to multiyear purchases from Mistral’s European Compute Units, a project intended to support 1GW of capacity by 2030.