Why the next wave of European Unicorns won’t come from Berlin, Paris or Amsterdam

There is a side street in Freiburg im Breisgau where, in 2024, two researchers knocked on the door of the local startup hub and asked if there was a desk free. There wasn’t. The hub was full. So Robin Rombach and Andreas Blattmann found a small office nearby, set up their laptops, and got to work.

Less than eighteen months later, their company – Black Forest Labs – closed a $300 million Series B at a valuation of $3.25 billion. Its image-generation model FLUX.1 had, in the words of the technical community, made every competitor look slow. Its customer list included Adobe, Canva, Microsoft, and Meta. And its team numbered roughly fifty people.

Fifty people. From a side street in Freiburg. This is the story the European startup ecosystem hasn’t quite processed yet.

The old logic made sense – until it didn’t

For the better part of two decades, the advice was simple: if you’re serious about building a startup in Europe, move to a hub. Berlin, Paris, Amsterdam, Stockholm, London – these cities offered the density of talent, capital, and networks that startups need to grow fast. The logic was sound. Hiring was easier. Investors were nearby. The coffee shops were full of people who had done it before.

The problem is that this logic was never really about geography. It was about access. Hubs concentrated resources that were scarce and unevenly distributed. Talent gravitated to cities because that’s where the jobs were. Investors clustered together because deal flow follows relationship networks. The geography was a side effect, not the cause.

That distinction matters enormously in 2026 – because AI has already dissolved much of the scarcity that made hub concentration necessary in the first place. Not “is starting to dissolve.” Has dissolved. The question is whether the funding infrastructure has processed that reality yet. It hasn’t.

What AI has actually changed – and what it hasn’t

The wave of AI tools that reshaped the startup world from 2022 onward is often discussed in terms of productivity: you can write code faster, produce content faster, handle customer support at scale. True. But by now, those gains are table stakes. Every founder has access to them. The deeper shift is structural, and it compounds over time.

Tasks that previously required specialist hires – product design, legal drafting, financial modelling, competitive research, even early sales – can now be handled by a solo founder with the right tools and enough domain knowledge. The minimum​ viable team for a software startup has dropped from ten to three. In some cases, to one.

Black Forest Labs is the extreme version of this thesis. Robin Rombach, Andreas Blattmann, and their co-founder Patrick Esser had been part of the research group of Björn Ommer – a lab that began at Heidelberg University and later moved to LMU Munich – which produced the foundational architecture behind Stable Diffusion.

When they left Stability AI in early 2024, they didn’t go to Berlin or London. They went to Freiburg – a university city of around 230,000 people in the southwest corner of Germany, better known for its cycling culture and the Black Forest on its doorstep than for any startup scene.

The choice was deliberate. Andreas Blattmann grew up in nearby Elzach. The region was home. And critically, it didn’t matter that it wasn’t a hub. Their competitive advantage wasn’t proximity to a VC network. It was a decade of accumulated research expertise that nobody else had – and a product, FLUX.1, that the market immediately recognised as the new standard.

When team size becomes optional, and domain knowledge is your moat, location becomes negotiable.

It’s worth being precise about what Black Forest Labs proves and what it doesn’t. Its moat was not AI-enabled leanness; it was a decade of rare research expertise that almost nobody else had. A sceptic could reasonably argue it would have succeeded anywhere – and that is exactly the point.

When the moat sits in the founders’ heads rather than in a city’s network, location stops being a constraint, and the team is free to optimise for what a smaller place does better: focus, low cost, and roots.

For founders in non-metropolitan Germany, Poland, Romania, or rural France, this isn’t a productivity story. It’s a structural levelling. The resources that used to require a Berlin postcode are now accessible from a broadband connection.

The structural advantages that hubs can’t replicate

Here’s the contrarian observation: non-metropolitan founders don’t just benefit from AI parity. In several important dimensions, they hold structural advantages that hub- based startups actively struggle to match.

The first is burn rate. A founding team operating out of Leipzig, Rostock, or Rzeszów can run lean in a way that’s genuinely difficult in Berlin or Paris. Office costs are lower. Salaries are lower – not because the talent is worse, but because the cost of living is different.

Black Forest Labs reached a $3.25 billion valuation with fifty​ employees; that ratio of valuation to headcount is possible in part because Freiburg is not San Francisco. A team running, say, 40% leaner on monthly burn can iterate longer, raise less, and survive the kind of slow-burn capital market that has defined Europe since 2022.

The second advantage is proximity to real problems. The German Mittelstand – the roughly 3.5 million small and medium-sized businesses that form the backbone of the country’s economy – is not headquartered in Berlin. It’s in Gütersloh, in Wolfsburg, in Memmingen.

Founders who grew up in these regions, who have family connections to manufacturing, logistics, trades, and agriculture, have a kind of problem intimacy that is genuinely hard to manufacture from a co-working space in Kreuzberg. They understand the pain before they write the pitch deck. BFL’s founders understood generative AI research at a level nobody else did. That’s domain knowledge from a different domain – but the principle is the same.

The third is talent that hasn’t been competed away. Smaller cities and rural regions have universities, polytechnics, and vocational schools producing capable graduates who often stay local. Freiburg has a strong university; Heidelberg, where BFL’s founders began their research, is not Berlin. The talent market in these cities is less liquid – which means it’s more accessible to early-stage companies without the brand recognition to win in saturated markets.

The funding gap is real – and it’s getting worse, not better

The honest counterargument is capital. Geographic concentration in European VC hasn’t narrowed – it’s deepened. London alone raised more venture funding in 2025 than the next twenty European cities combined.

For a founder in Erfurt or Gdańsk, the practical reality of accessing Series A capital is still substantially harder than for their counterpart in Mitte or Shoreditch. The data does not support optimism here.

It’s worth noting that Black Forest Labs navigated this partly by accepting US capital – Andreessen Horowitz led the Seed round, and Salesforce Ventures and AMP co-led the Series B. The company is formally registered in the US as well as Germany.

A sceptic could argue this makes BFL a US-backed company with a German office, not a non-metropolitan European success story. That reading underestimates the significance of where the company was built and where its talent sits. But it points to a real structural problem: European non-hub founders often still need to reach outside Europe to access growth capital at scale.

What is shifting, slowly, is the source structure of capital within Europe.

In 2025, French and German VC firms dominated the rankings of new funds raised, taking all but three of the top 10 spots – a notable contrast to 2024, when London-based funds held eight of the top ten. This matters because a broader geographic distribution of fund formation tends, over time, to pull investment sourcing away from a single gravitational centre.

It is a weak signal, but it is a real one.

The more durable structural shift is the rise of national development banks and EU cohesion instruments as funders of non-hub ecosystems. These vehicles don’t replace venture capital, but they provide the bridge that lets a non-metropolitan startup get to​ the point where a VC conversation makes sense.

Closing that gap remains the single most important policy lever available to the European startup ecosystem.

What the ecosystem needs to do

The startup ecosystem – investors, accelerators, policymakers, media – still largely operates on a mental map drawn in 2015. The assumption baked into most evaluation frameworks is that serious startups cluster in serious cities. Black Forest Labs is evidence that this assumption is wrong.

But BFL is an extreme case – world-class researchers with a decade of accumulated IP. The question is what happens to the broader cohort of non-metropolitan founders who are less exceptional, but still structurally better positioned than the ecosystem gives them credit for.

A few concrete shifts would help.

Accelerators should move from city-centric to sector-centric models, measuring impact by outcome rather than alumni count, and deliberately recruiting from non- metropolitan regions where talent is systematically underrepresented.

The most valuable thing an accelerator can do in 2026 is not provide office space in a co- working hub – it’s connect a founder in Bremerhaven to a customer in Stuttgart and a fund in Munich.

Investors should weight regional distribution as a positive signal. A founding team with deep roots in a market they’re serving is not a red flag; it’s often a moat. The pandemic-era lesson that geography matters less to deal quality than assumed has been quietly absorbed by many European funds – but it hasn’t yet translated into systematic sourcing changes.

EU and national funding instruments should close the equity gap for non-metropolitan founders, many of whom are currently channelled exclusively toward grant programmes that don’t build the same institutional muscle as equity-backed growth.

Grants extend runway. They don’t build the investor relationship, the governance discipline, or the scaling muscle that comes with equity. Both are needed.

The thesis, stated plainly

The combination of AI-enabled execution and structural regional advantages is producing a cohort of European founders who are better capitalised on time, better connected to real customer pain, and less exposed to the cost pressures that kill early-stage companies.

Black Forest Labs didn’t need Berlin. It needed a decade of research, a broadband connection, and a product that was simply better than everything else. The Freiburg address wasn’t a handicap. In a perverse way, it might have been an asset – fewer distractions, lower burn, a tight team that wasn’t competing with fifty other well-funded AI startups for the same senior engineers.

In 2025, new unicorns emerged from eleven different European countries. The next one might be registered in a city you’ve never heard of. That’s not a bug in the European startup ecosystem. In 2026, it might be exactly why it wins.​

The European startup ecosystem has spent two decades trying to build hubs that look like Silicon Valley. The more interesting question now is what gets built by the founders who were never invited to the hub – and no longer need to be.

The post Why the next wave of European Unicorns won’t come from Berlin, Paris or Amsterdam appeared first on EU-Startups.