Europe's Startup Exodus: Why the US Wins at Scale

Jan de Vries · 2026-09-14

I don't believe the neat story of a mass exodus of European startups. The numbers just don't support it. Most companies stay where they were founded, and those that do move rarely make that decision overnight. The real question is: what happens to the companies that begin to scale? ### A company doesn't relocate in a day The European Commission's Joint Research Centre estimates that only 3.3–4.3% of European venture-backed companies relocate fully or partially. Among comparable companies without venture capital, that figure drops to 0.3–0.5%. That's not a mass departure. Yet 97% of the companies that do relocate do so partially. They keep operations in their country of origin: a lab, engineers, manufacturing, or part of the team. The holding company, headquarters, leadership, fundraising, or commercial function might move elsewhere. When that happens, the US is the main destination. But "relocation" isn't quite the right word. The center of gravity shifts gradually—from one function to another, not in a single move. ### The more successful the company, the more expensive the next step The picture gets sharper for companies at the scaling stage. According to the 2024 Draghi report, as cited in the European Investment Bank's 2026 analysis, around 10% of European scale-ups relocate abroad, and roughly 85% of them choose the US. The report also notes that close to 30% of European unicorns founded between 2008 and 2021 moved their headquarters abroad, predominantly to the US. These figures cover different groups and periods, but they point to the same mechanism: Europe risks losing a disproportionate share of the value created by the businesses that grow into large international companies. In biotech, the gap is especially visible. An EIC report found that between 2019 and 2025, 66 of the 67 EU biotech companies that went public chose exchanges outside the EU. That doesn't mean 66 companies closed their European offices. It means the market for their IPOs, future liquidity, and next stage of capital was almost always located in another system. ### After the laboratory comes a different economy Europe is strong at the early stage. It has universities, research centers, Horizon Europe, the European Innovation Council, and national funding programs. When a company enters the next stage of its economic development, its need for money doesn't disappear. What changes is the job that money must do. Early capital funds research and technology development. The next step is to scale manufacturing, navigate regulation, build international sales, secure hospital procurement, and hire executives who have already brought a product to market. A company therefore needs more than capital. It needs capital connected to a market, specialist networks, and execution experience. The European Central Bank estimates the total stock of venture capital at around $1.01 trillion in the US, compared with $163 billion in the EU. This is not annual investment or dry powder; it's an estimate of the overall size of each system. The gap matters most at later stages, when the sums become larger and mistakes more expensive and difficult to reverse. A founder isn't simply choosing a check. They're choosing the system in which the company is less likely to make a costly mistake. ### The problem isn't consultants. It's the cost of navigation European funding programs are surrounded by consultants, grant writers, and intermediaries. A good consultant saves time and explains rules that a founder may be encountering for the first time. Across Horizon Europe, 17% of applicants used external consultants or experts; among EIC Accelerator applicants, the share reached 67%. The cost also varied considerably. The median consultancy fee was $8,200 for a consortium proposal, $2,200 for a single-beneficiary proposal, and $13,100 for an EIC Accelerator application. So what's the takeaway? It's not that Europe can't build great companies. It's that scaling them often means plugging into a different ecosystem. And that's a choice founders make step by step, not all at once.