Europe's AI Dependence: Why Startups Risk Renting Back Their Own Knowledge

Jan de Vries · 2026-10-13

AI is quietly reshaping how European businesses operate. What started as a tool for analyzing data and handling basic admin tasks is now woven into the daily workflows of companies across the continent. But here's the catch: most of that AI comes from somewhere else. The most advanced AI models and infrastructure are concentrated in the US and China. That gives those countries enormous influence over a technology that's becoming as essential as electricity. And with both nations pushing their own AI ecosystems, the rivalry is spilling into geopolitics. Recent reports suggest the US is preparing to tell dozens of countries to pick sides between American and Chinese AI. China, meanwhile, is calling for respect of "digital sovereignty." AI sovereignty isn't a theoretical debate anymore. It's happening right now, and Europe risks becoming a customer of whichever superpower controls the intelligence it depends on. ### Europe doesn't need self-sufficiency, but it needs choice The vulnerability was on full display in June when the US ordered Anthropic to restrict foreign nationals' access to its most advanced models. That sent a shiver through European boardrooms. How much autonomy will Europe have if this divide widens? Europe should be able to use the best models, wherever they come from. But it can't build its economy on imported AI that might get cut off. It doesn't need to be completely self-sufficient—it just needs enough options to switch suppliers and keep viable alternatives alive. That choice matters more as AI becomes embedded in everyday business operations. In July, the European Commission launched a call for up to seven AI Gigafactories, backed by up to $10.8 billion in EU and national funding, aiming to unlock at least $21.6 billion in private investment. It's a start. But sovereignty isn't just about the models. It's also about the knowledge businesses create through AI. ### The hidden risk: renting back your own knowledge Every time an employee uses AI, they're documenting how their work gets done—often for the first time. Companies that hold onto that knowledge can use it to fuel their AI transformation. But many risk handing it over to their vendors, teaching them how to commoditize their business, and eventually renting back their own know-how. The consequences go beyond technology. AI will determine which companies become most productive and which industries adapt fastest. If European companies import these capabilities from foreign entities, the value they create will likely flow outside Europe too. This creates a different kind of dependency. A business might switch model providers, but separating from a system that has spent years learning its processes and organizational knowledge is much harder. The longer a company builds its operations around external AI providers, the more entrenched that dependency becomes. > "The companies that hold onto that knowledge can use it to fuel their AI transformation. But many risk handing it to their vendors, teaching them how to commoditise their businesses, and will end up renting back their own know-how." For startups and scale-ups, this is especially critical. Their advantage often lies in specialized knowledge—their understanding of a problem and the distinctive way they work. If that knowledge gets trapped inside an external platform, they risk giving away part of what makes their business valuable. This doesn't mean European businesses should stop using foreign AI. That's unrealistic and would pose its own threats. But they need to be strategic. They need to keep control of their data, their processes, and their organizational knowledge. Otherwise, they'll find themselves renting back the very things that make them unique.