There is a version of Europe’s tech story that reads really well right now. Deep tech funding hitting records. Mistral going toe to toe with OpenAI. Lovable pulling in hundreds of millions from Silicon Valley funds that spent years ignoring anything with a European postcode. University spinouts from Cambridge, ETH Zurich and the Nordic research institutions finally getting the kind of backing they deserved a decade ago. Read the reports from Atomico or Dealroom and you come away feeling like the European startup moment has genuinely arrived.
Then you talk to the founders actually building these companies, and a different picture emerges. There is growing conversation in policy circles about a common corporate vehicle referred to as EU Inc which would let tech founders incorporate once and operate across multiple countries without repeating the whole legal process from scratch in each one. It is a genuinely good idea, and the fact that it is being discussed seriously at all is progress worth acknowledging. But it is not here yet. It is still being shaped, still being argued over in rooms that founders do not have access to, and the people building companies today are dealing with the world as it actually is, which is considerably messier.
What is also in motion is the reform push that regulators have been calling the 28th regime, a proposed optional legal layer that would let tech startups choose to operate under a single unified European framework rather than trying to stitch together compliance across two dozen different national systems. The logic is sound. A software company selling to customers in six countries should not need to employ lawyers in six countries just to understand its own obligations. Whether this translates into something real and usable for an early-stage founder is a question that nobody can fully answer yet. But at least the diagnosis is correct, which is further than European policymakers have historically been willing to go.
What Founders Are Actually Dealing With
While the policy discussions continue, the decisions founders have to make are national, specific and consequential in ways that most early-stage advice does not properly prepare people for.
Take Germany. It is the biggest tech market in continental Europe and a logical target for almost any B2B software or data company with serious scale ambitions. But setting up a GmbH, which is the structure you actually want if you are going to be taken seriously by German banks, German enterprise clients and German institutional investors, requires 25,000 euros of minimum share capital, a notary, and a process that typically takes several weeks to complete. For a small founding team trying to move fast, a few weeks is not a trivial cost. There is a cheaper entry point called the UG, which can be started with almost no capital, but it comes with restrictions on distributing profits and tends to get noticed by the kind of counterparties you most want to impress. A lot of founders use it as a temporary structure and convert later. It works, but it adds a layer of process at exactly the point when process is the last thing you want.
France is a different story and, in some ways, a more encouraging one. Paris has built a genuine AI cluster over the past few years, and Mistral’s rise has done something important for the ecosystem’s psychology: it has shown that a European AI lab can compete on the world stage rather than just being acquired by one of the American giants before it has the chance. The SAS, which is the corporate structure most tech founders use in France, is reasonably flexible and allows for the kind of governance arrangements that early-stage investors expect to see. Formation is slower than in the UK or Estonia, but the process has become more navigable as more of it has moved online. The French Tech initiative has done genuine work here, even if it does not always get credit for it.
Estonia: The One Everyone Asks About
Estonia comes up in almost every conversation about European startup formation, and the interest is not irrational. The OÜ can be set up and run entirely online. The e-Residency programme makes it accessible to founders who have no prior connection to the country. The tax treatment is legitimately different from most places: profits are not taxed when they are earned, only when they leave the company as dividends. For a startup that is growing and reinvesting heavily, that deferral is real and meaningful over a period of years.
The thing that tends to get left out of the enthusiastic writeups is what happens when you look more carefully at the substance question. If your company is an Estonian OÜ but every decision is being made in a flat in Berlin or a co-working space in Athens, most European countries have rules that look at where a company is actually managed from, not just where it is registered. Those rules can make your Estonian company tax resident in your home country regardless of what it says on the incorporation certificate. That does not make the Estonian structure useless. Used properly, by founders with real operational ties to the country or genuinely distributed teams, it holds up well. But it is not a loophole, and founders who treat it like one tend to get an unwelcome education eventually.
The UK, Which Is Now Its Own Separate Calculation
Before Brexit this would have been a simpler section to write. A UK limited company was the obvious default for a lot of European tech founders, particularly those building in English and targeting international markets. Companies House is still one of the most efficient registries in the world. You can have a company set up in hours for almost no money, and the structure is instantly recognisable to investors, banks and clients almost everywhere.
What changed is the implied access. A UK entity no longer gives you any automatic right to operate across continental European markets, which means founders who want to build seriously across both the UK and Europe are increasingly maintaining two separate legal structures, one on each side. That costs money, takes time, and adds administrative overhead at the stage when you can least afford it. It is not insurmountable, but it is an extra layer that did not used to exist, and pretending otherwise does not help anyone.
The Part That Actually Keeps Founders Up at Night
The honest version of this conversation is that the legal and administrative burden of starting a tech company in Europe is not just a cost in euros. It is a cost in attention, and attention is the thing early-stage founders have least of.
A machine learning engineer or a data scientist who is also trying to be a founder is someone who is, every day, choosing between building the product and dealing with everything else. When the everything else includes understanding the difference between authorised and paid-in share capital, working out what equity instrument is appropriate for the first three advisers, figuring out whether you need a local director in a jurisdiction you are trying to sell into, and getting to grips with VAT registration rules that vary by country and by product type, the cognitive load is genuinely significant. None of it is impossible, but all of it requires either expensive professional help or hours of self-education that could have gone into the actual company.
And it compounds. A startup operating across three European countries is not dealing with three times the administrative burden of one operating in a single country. The interactions between different national frameworks create their own complications that no single adviser in any one country is equipped to fully navigate. You end up needing specialists in multiple places, coordinating between them, and hoping they are all working from the same understanding of your structure.
Why This Matters Beyond the Paperwork
There is a talent dimension to this that does not get talked about enough. A strong senior engineer or an experienced product leader evaluating an offer from a European startup is going to have questions about equity. How is the cap table structured? What kind of option scheme is in place? How does the vesting work? How has the company thought about exits?
If the answer involves a corporate structure that the engineer’s accountant has never seen, a share class that does not have a clean English-language explanation, or an option scheme that was set up quickly and cheaply and has never been properly reviewed, that affects the decision. Maybe not on its own, but combined with everything else, it matters. The founders who treat legal structure as something to sort out fast and cheaply in week one, and fix later, often find that later comes at a moment when fixing it is significantly more expensive than getting it right initially would have been.
Europe is producing tech companies with genuinely global ambitions in a way that was not true ten years ago. The foundation those companies are built on deserves more attention than it usually gets.