I want to build European software for Europeans. On European infrastructure. Independent. And right now, that means fighting through every layer myself—because the stack nobody assembled for you does not exist yet.

I am building two products in Germany, both for the European market. And finding the right infrastructure is like furnishing a house where IKEA does not exist—every piece comes from a different shop, in a different city, with a different return policy.

In the US, you spin up a startup and the ecosystem hands you everything: Vercel for hosting, Auth0 for login, Stripe for payments, SendGrid for email, Sentry for errors. It works. It integrates. You ship in a weekend.

In Europe? You start Googling.

What I Found

Authentication. The European alternative is Ory, based in Munich. Open source, self-hostable, genuinely good software. But I found Ory after three days of research. An American founder finds Auth0 in three minutes because it is the default answer on every Stack Overflow thread, every tutorial, every “How to build a SaaS” blog post.

Hosting. In the US, you do not even need to think about servers. Vercel, Cloudflare Pages, Railway—generous free tiers, push to Git, app is live. Hetzner is phenomenal and cheaper than any US platform at scale. But in the US, deploying is a solved problem. In Europe, deploying is a skill.

Payments. Stripe is everywhere. Mollie is Dutch, Adyen is Dutch—both excellent. But every payment integration guide assumes Stripe. Every SaaS boilerplate ships with Stripe. If you want Mollie, you are on your own.

AI Models. This is the most honest part. There is no European foundation model today that matches Claude or GPT-4. Mistral in Paris is closest but holds roughly 2% of the global LLM market. Aleph Alpha in Heidelberg raised half a billion euros to build “Europe’s OpenAI”—and quit the race in 2024.

I use Anthropic’s Claude. American company. I am not going to pretend otherwise.

The Real Problem Is Not the Components

Every European alternative I listed works. But working is not enough. The US tools do not just work—they connect. Click, GitHub, deploy. Stripe gives you hosted checkout, a billing portal, subscription management. Mollie gives you an API and says: build the rest. Vercel gives you preview deploys on every PR. Hetzner gives you a VPS and SSH access.

The European components are capable. But they do not have the same developer experience, the same polish, the same automatic integration. And nobody assembled them into a stack. There is no “European SaaS Starter Kit.” No opinionated boilerplate that wires Hetzner + Ory + Mollie together and says: here, deploy this and you are sovereign.

In German we call this kind of invisible work Kleinarbeit—the small, unglamorous work that nobody sees but everything depends on. Building European software means doing a lot of Kleinarbeit that American founders never have to think about.

What Europe Actually Needs

Not more regulations. Not another EU-funded “cloud initiative” that produces PDFs instead of products.

A zero-cost experimentation layer. In the US, a founder validates ten ideas spending zero dollars. In Europe, every idea has a minimum viable cost. Zero and non-zero are different categories. Zero means try everything. Non-zero means evaluate first. Across an entire continent of potential founders, that difference compounds.

An opinionated default stack. The European equivalent of create-t3-app—a starter kit that makes the sovereign choice the easy choice.

Founders who document the path. That is what I am doing. Not because I have it figured out, but because the next person building European software should not need three days to find their auth provider.

This Is Part One

DHH left AWS and saved $10 million over five years. He published every number, every decision, the entire stack. He called the cloud vendors “merchants of complexity” who convinced an industry it could not run its own computers. He proved it by doing it, not by writing position papers.

I am doing the same thing for the European stack. Not at DHH’s scale. But with real products, real users, and real switching costs. This is the first post. There will be more—on the auth decisions, the payment migration, the things that break and the things that turn out easier than expected. A field journal called Rohbau. Because that is what the European software stack is right now: a shell with solid walls, waiting for someone to do the interior.

I am doing the interior. Follow along.

European software sovereignty is not about rejecting American technology. It is about the Kleinarbeit of assembling a stack that nobody assembled for you—and being honest about where the alternatives hold up and where they do not.

Sources:

Paste into LinkedIn articles, Notion, or any rich text editor.