Founders talk about “going to the US” as if chasing easier money: faster yeses, bigger cheques, fewer questions about the plumbing. They’re acting on it: Index Ventures found 64% of companies now expand to the US at pre-seed or seed stage, up from 33% between 2015 and 2019.
The gap is real, but misunderstood. I spent years underwriting deals before advising founders, and “easier” is rarely how capital behaves. Two markets are pricing the same digital and AI-driven risk differently. Seen that way, the decision is arithmetic, not ambition.
Start with the market, not the money
A US software business sells into one domestic market: one language, one currency, one federal regulatory perimeter. Sign a customer in Ohio and the same product works in Texas. Europe doesn’t: Mario Draghi’s report on EU competitiveness counted more than 270 digital regulators across member states and roughly 100 laws aimed at tech. Every new country means a new language, a new payment habit, often a new licence and supervisor.
This is the part founders skip, and it’s what I keep coming back to with clients. A US investor paying up for potential isn’t reckless: they’re pricing a market where a good product reaches millions without renegotiating its right to operate. A European investor asking for traction first is pricing one where growth is won country by country, and the second costs almost as much as the first. Draghi found Europe produced 147 unicorns between 2008 and 2021, 40 of which relocated abroad, mostly to the US; the Commission put it plainer still in May 2025: only 8% of global scaleups are based in Europe.
Founders often collapse two decisions into one. Wanting American customers isn’t the same as needing American investors: you can sell into the US from London, and plenty of companies do. One is a commercial choice, the other a financing choice, and they rarely need making in the same year.
The US pays for ambition, Europe pays for traction
In Q2 2025, just under 70% of global venture investment went to US companies, per CB Insights. IMF research covering 2013–2023 puts US venture fundraising at roughly $924bn, against about $130bn across the EU. Valuations follow: reading State of European Tech 2024 data, Equidam puts the European discount at 29–52% depending on stage, widest at seed. That number does the most work: at seed there’s rarely much of a business to underwrite, so paying roughly double is a bet on how large the thing might become, not diligence. AI has sharpened the pattern: CB Insights reported AI companies raised $47.3bn globally in Q2 2025, with $39.7bn going to US firms.
European capital wants more proof and pays less per unit of ambition, which is rational given what it’s underwriting. European venture funds have matched or modestly outperformed North American funds over ten years, per Invest Europe and Cambridge Associates, and a seed-stage company is no less likely to reach a billion-dollar valuation there than in America. The gap is in how many get funded: five years in, an American startup is 40% likelier to have raised venture money.
Fintech and AI aren’t the same trade
In fintech, the regulatory perimeter is the product. UK firms lost the right to serve EU customers on a domestic licence when passporting ended in December 2020; Wise saw it coming, applying to Belgium’s National Bank for a payment licence in 2019 to keep its EU customers, licensing work a founder would hand over by relocating.
AI has no equivalent gate: a model ships globally on day one, needing compute and capital, not permission. That’s where Europe is thinnest, attracting about 6% of global AI funding to the US’s 61%, though European AI application companies now take in 66 cents per dollar raised by US counterparts, up from a tenth a decade ago, per Accel. Building models, the constraint is capital at scale; building applications, it’s domain knowledge, which Europe isn’t short of.
The real constraint is growth capital
Early-stage Europe is stronger than its reputation; the shortage comes later. European pension funds hold around $9 trillion and allocate about 0.01% of it to European venture, a gap Atomico sizes at $375bn in missed growth-stage funding. When a company needs hundreds of millions for infrastructure, the domestic pool runs short, and founders look west. Wise makes the point twice: it kept its European customers, then in July 2025 its shareholders voted to move the primary listing to a US exchange.
Three questions before you raise
Ambition or traction? US investors pay for what you could become; European investors price what you’ve built.
Customers or capital? A market problem is commercial: sell in, hire, localise. None of that requires touching your cap table. A capital problem is financing, and only then does investor location matter.
Growth capital, now or later? Flipping into a US structure is a financing decision, not a relocation. Reactive, it’s costly; timed ahead of a justified round, it’s just plumbing.
The money isn’t easier in the US. It’s pointed at a different risk, in a market built for a different kind of growth. The question isn’t where cheques are looser; it’s which risk you’re asking investors to price.
Evaluating risk is what I do for a living: regulation, ownership, governance, timing. A risk that’s a footnote in a big balance sheet can be fatal in a young one, which is exactly why I look at early-stage fintech and digital companies in depth, even when the cheques are small.
Bio:
Michelle Luan is an Investment Banking Associate in M&A and a Senior Executive, specializing in risk evaluation across technology-enabled businesses. She works with founders to refine strategy, prepare for fundraising, and de-risk their path to capital. She has advised on large capital raises and M&A transactions (c. $500m–$5bn) for companies operating in fintech infrastructure, digital platforms, data-driven mobility, and other technology-led businesses.
The views expressed are the author’s own and drawn from general professional experience. They do not represent any current or former employer, and no confidential or client-specific information is referenced.
