Europe's innovation gap with America widens despite Draghi warning

When Mario Draghi presented his European Commission report on competitiveness in September 2024, the numbers were stark enough to silence a room. Over the previous half-century, not a single European company created from scratch had reached a market capitalisation of €100bn. In the same period, six American firms had sailed past $1 trillion. Since 2013, the United States had spawned 137 venture capital funds larger than $1bn; the European Union had managed eleven. And in the sector now defining the next economic era, artificial intelligence, US startups absorbed 61% of global funding in 2024. The EU's share was 6%.

The EU's population exceeds America's by more than 100 million people. The disparity is not a matter of market size. It is structural.

A regulatory thicket no single market can clear

Adriana Hoyos, an adjunct economics professor at IE University in Madrid, describes the European regulatory architecture as a "very slow elephant". "You have European regulation, then you have the national regulations, then you have state regulations with all these completely different ways of behaving," she said. The Draghi report counted roughly 100 tech-focused laws and more than 270 regulatory authorities with a say over digital networks across the bloc, from telecommunications to data protection.

The AI Act, adopted in 2024, exemplifies the compliance burden. It categorises models by risk level and imposes escalating obligations. While the safety logic is sound, particularly for law enforcement, hiring and education, the cost of compliance falls disproportionately on young companies without legal departments. A French founder at Mistral's AI Now Summit in Paris this May put it bluntly: "We spend 20% of our engineering time on regulatory mapping. That is 20% we are not spending on the model."

Fragmentation compounds the problem. A startup scaling from Berlin to Milan to Warsaw encounters three tax codes, three labour regimes, three sets of corporate governance rules. The single market exists for goods; for digital services and capital, it remains a mosaic.

Capital markets that do not scale

Josh Lerner, who teaches investment banking at Harvard Business School, has documented the venture capital gap in both volume and returns. In the United States, half of all companies that have gone public in the past two decades were venture-backed. Nearly 90% of R&D spending by young public firms comes from venture-backed companies. "All the innovation in the United States being done by dynamic, young, recently public companies is basically being done by venture-capital-backed firms," Lerner said.

Europe lacks a continental exit market. Nasdaq provides a deep, liquid venue for US tech IPOs. In the late 1990s, a group of investors launched EASDAQ as a pan-European equivalent. Competing regional exchanges fragmented liquidity, and EASDAQ shut down in 2003. No successor has emerged. The consequence is visible in the data: between 2008 and 2021, nearly 30% of European startups that reached a $1bn valuation moved their headquarters abroad, overwhelmingly to the United States, according to the Draghi report.

Tax regimes that push founders out

The capital problem is reinforced by taxation. Lerner cites research by Christine Blandhol, a doctoral candidate, showing that an increase in Norway's wealth tax rate coincided with a jump in out-migration among affected households from 0.2% to 2%. Forty percent of departing households were active firm owners. The wealth trajectory diverges sharply: between 2014 and 2024, the number of US millionaires rose 78%. In Germany the increase was 10%; in France, 7%. The United Kingdom saw a 9% decline.

These are not abstract aggregates. They represent founders deciding whether to build in Munich or move to Palo Alto. The decision is increasingly made before the first funding round.

The brain drain measured in doctorates

Hoyos notes that roughly 75% of European PhD students at American universities remain in the United States at least five years after graduation. "The ones that are interested in technology, basically all of them want to go to the US to work or to study for advanced degrees," she said. The US university system functions as a talent magnet; the European system functions as a feeder.

This is not new. What is new is the compounding effect: AI research clusters around compute and capital, both concentrated in the United States. European labs produce papers; American labs ship products.

Work culture and the ambition question

Nicolai Tangen, chief executive of Norway's sovereign wealth fund, put it bluntly in a 2024 Financial Times interview: "We are not very ambitious. I should be careful about talking about work-life balance, but the Americans just work harder." The numbers bear him out. The EU Working Time Directive guarantees at least four weeks of paid annual leave, non-monetisable. In the United States, 31% of workers have no paid leave at all, and the average is eleven days, just over half the European minimum.

Whether longer hours translate into higher productivity is debated. What is not debated is that the intensity of early-stage startup life, the 80-hour weeks, the iterative speed, is culturally accepted in Silicon Valley in a way it is not in Berlin or Paris. That cultural permission matters when the competition is measured in model training runs per quarter.

The Trump variable: opportunity or distraction?

The second Trump administration has introduced policies that could, perversely, create space for Europe. The Wall Street Journal reported a proposed $100,000 fee for foreign graduates to work in the United States, after a similar fee for H-1B sponsors was struck down in June. A July rule eliminated the "duration of status" provision that allowed F-1 visa holders to remain for the length of their academic programme without a fixed expiry.

Lerner recalls a joke he made to Canadian officials: the first Trump administration did more for Canadian venture capital than thirty years of Canadian policy. The logic is simple, restrict the world's talent magnet, and talent looks elsewhere. Europe could be that elsewhere. But only if it is ready.

Defence spending offers another lever. The European Commission's Readiness 2030 plan envisages up to €800bn in additional defence expenditure over five years, including mobilisation of private capital. Dual-use technologies, drones, satellite communications, cyber, could seed a new generation of European deep-tech champions. The Commission's own communications frame this as an industrial policy opportunity.

What a turnaround would require

Hoyos argues that incremental tweaks will not close the gap. "Europe is clearly lost in the technological world," she said. "They don't see the priority, they don't see how fast this is going, and they think it's something that's optional." A mindset shift would mean treating AI and quantum not as regulatory targets but as strategic infrastructures, comparable to energy or transport in the 20th century.

Concretely, that implies a single digital incorporation regime, a consolidated capital markets union with a genuine European growth exchange, tax coordination that stops member states bidding against each other for mobile founders, and a visa framework that makes hiring a non-EU AI researcher as frictionless as hiring a German one. None of these are technically impossible. All require political decisions that have been deferred for a decade.