Why Europe is falling behind in AI — and how to catch up

AI: Why Europe is falling behind, and how it can catch up July 16, 2026Philippe Aghion is an expert in creative destruction. That's the term economists use when innovations disrupt whole industries and create new ones. Think steam engine, assembly line production and the IT revolution. Or now, artificial intelligence (AI). For his groundbreaking research on how technological innovation can drive long-term economic growth, Aghion, who is French, was awarded the Nobel Prize for economics in 2025, together with Canadian Peter Howitt and Israeli-American Joel Mokyr. "More than ever, AI will bring creative destruction," Aghion told DW on the sidelines of this year's Brussels Economic Forum. "There will be job destruction. And there will be job creation. If you manage creative destruction the right way, it's an engine of social promotion." The importance of embracing change To illustrate the benefits of creative destruction, Aghion points to the difference in economic growth in the European Union (EU) and the US. For decades, they developed in lockstep. But when the IT revolution happened — the personal computer, the internet and everything that goes with it — the US fully embraced the changes. Europe not so much. As a result, the gap widened. Today, per capita gross domestic product in the US is nearly twice that of the EU. With AI, Europe is at risk of repeating the mistake all over again. "We need to wake up," Aghion warned. The US and, to a lesser degree, China are far ahead when it comes to AI development. Firms like OpenAI, Anthropic, Google, Meta, Microsoft and many others are investing billions into the development of large language models (LLM) like ChatGPT, Gemini, Claude or DeepSeek, a Chinese model, and the computing power needed for these models. Finding strength in EU data protection Europe produces excellent research but has historically struggled to scale startups and attract the same level of venture capital as the US. European models like the French Mistral are the exception to the rule. While it might be hard or even impossible for Europeans to catch up in the area of LLMs, Aghion argues they should bring AI to their traditional strengths like health. "We have fantastic health data in Europe, much better than in the US. And there will be fantastic AI health applications. We can develop a lot of specialized AI." And if done well, the European penchant for regulation could turn out to be a strength, for example in data protection and privacy. "I believe there will be a demand for a more ethical AI, an AI that avoids some dangers," he said. "People want to be protected. To have some regulations will make us even more attractive." How Europe could catch up in AI Some business leaders agree. Thomas Saueressig, vice chair of the Board of Directors of SAP, Europe's biggest software company, wants to see AI used outside the confines of LLMs, to put traditional European industries "on steroids." "Take our capabilities in manufacturing, for example: We could use physical AI and bring it to the next level," Saueressig said in DW's business podcast "The Dip." "We need to use this technology to leapfrog our existing model into the future. In uncertain times, the biggest risk is not to take a risk." But to do that, Europeans need to do their homework, insists Nobel laureate Aghion. "First, we need to create a better research environment, with more long-term financing of research," he said. "We also need more venture capital, and more pension funds." Why Europe needs more venture capital European startups, whether they deal with AI or not, generally have a hard time to secure financing. Banks are often reluctant to finance startups because many young companies fail and lack collateral. "It's different for specialized investors," also called venture capital firms or angel investors, said Marlene Schörner, a researcher on financial markets in the EU at the Jacques Delors Centre in Berlin. "They don't mind that many startups fail because they really profit from the successful ones they invested in. We need more of these investors." One way of attracting more venture capital is to harmonize the rules within the European Union. "Bankruptcy laws are very important for investors. They want to know whether or not they can get their money back if a business fails. But these laws vary a lot between EU member states," Schörner told DW's German-language business podcast "Wirtschaft im Gespräch." Can Europe's savings fund AI innovation? Another way to inject some life into Europe's startup scene is to harness the continent's huge private savings. "Europe sits on about €12 trillion [$13.6 trillion] in household savings," Luxembourg's prime minister, Luc Frieden, said during a speech at the Brussels Economic Forum in May. "But most of that money flows into real estate and government bonds, rather than investments in startups." To change that, the European Commission has launched the Savings and Investment Union, which is basically a rebranding of the Capital Markets Union, explains Marlene Schörner. "The goal of integrating European capital markets has been around for a very long time. However, this effort has not been very successful so far," she said. In the past, countries were reluctant to give up control of their financial markets. The hope is that recent changes in geopolitics, the discussion about whether the US is still a reliable partner, and the need for European self-reliance will put enough pressure on the 27 EU members to put their differences aside and build a common financial market. A European DARPA for AI Philippe Aghion believes research and financing could be much improved by bringing in the military. "We need something like DARPA," he said, referring to the Defense Advanced Research Projects Agency, a US institution charged with developing technologies for military use. DARPA was created in 1958 and played an important role in technologies such as the internet and GPS. Since Aghion doubts that all EU members would support the idea of a European DARPA, he suggests France and Germany take the lead. "We could have a French-German DARPA, and if other countries want they could join." The Nobel laureate describes himself as a "fighting optimist" when it comes to AI. However, he also has words of caution for European politicians. In order to cope with the destructive potential of AI, Europe needs to focus on education. This also means retraining those who lose their jobs. In countries without good social security and education systems, "AI will create a lot of unhappiness and frustration, and there's a huge risk that they go into populism for a long time," Aghion said. Edited by: Tim Rooks