The Economics Nobody Talks About
Software has an unusual cost curve. The first copy is expensive — engineers, design, testing, the lot. The second copy costs almost nothing. The millionth copy costs nothing either. Marginal cost of distribution is effectively zero once the work is done. With AI in the loop, even the first copy is getting cheaper.
This should mean software is dirt cheap. Instead most organisations pay more for it every year. How? SaaS. The subscription model is a coordination workaround for the zero-marginal-cost problem. If you can’t charge per copy, you charge per seat, per month, forever. The vendor captures the rent that the economics otherwise wouldn’t allow.
It works because coordination is hard. A thousand municipalities all need a citizen-reporting tool. Organising them to share one is painful. A SaaS vendor solves that coordination by being the single intermediary everyone pays. The vendor’s actual innovation is usually not the software. It’s the billing system.
Europe Is Starting to Notice
Something meaningful is happening in European public sector IT right now. In April 2026 France announced it is moving 2.5 million government computers off Windows to Linux, citing the need to “regain control of our digital destiny.” That came months after Paris replaced Microsoft Teams with Visio, a French tool built on the open source Jitsi.
The German state of Schleswig-Holstein is 80% done migrating its workstations from Microsoft 365 to LibreOffice, and is now saving €15 million a year in licence fees. Germany also runs ZenDiS, the Centre for Digital Sovereignty, which ships openDesk — already used by roughly 100,000 public-sector workers, with contracts for 60,000 more in the pipeline. Denmark’s Ministry of Digital Affairs has joined the LibreOffice wave. The European Commission puts the EU open source market at around €65 billion.
This isn’t arriving gradually. It’s arriving now.
But Each Country Is Doing It Alone
The pattern in every story above is the same. A country, a state, a ministry decides to quit Microsoft, then sets about building or commissioning their own stack. France has Visio. Germany has openDesk. Denmark has its own migration. Each one is solving roughly the same problem with a different project, paid for out of a different budget, maintained by a different team.
That’s better than the Microsoft status quo. But it misses most of the benefit. The whole point of open source is that the second copy is free. If France and Germany both need a videoconferencing tool, they should both use and fund the same tool. If every Dutch municipality needs a citizen-reporting platform, they shouldn’t each buy one — they should co-fund Signalen, which several already do. eVaka in Finland, Decidim in Spain: the proof of concept exists. What doesn’t exist is a way to do it at scale, across borders, for the long tail of software every public body already uses.
A Protocol, Not a Body
The obvious answer — a central European body that collects contributions and decides what to fund — would re-create the intermediary problem. It would also be slow, politically captured, and wrong about which projects matter. The better answer is decentralised. Every organisation keeps deciding which projects to back. They just all agree to do it.
The coordination lives in the commitment itself. A public charter that organisations sign: short, principled, versioned. Something like “we will not spend on proprietary software where a viable commons-funded alternative exists, and we will publish annually what we spend on software, broken down by proprietary and commons-funded.” A registry that maintains the list of signatories and their reports — lightweight, not a gatekeeper, just a ledger. Commitment-gated access, where software released under the scheme’s licence is free to use for signatories and non-signatories either sign or go buy SaaS. And downstream propagation built into the licence itself, so projects receiving funds commit to passing a share onward to their own upstream dependencies. Funding flows along the real dependency graph, not just to the visible top of the stack.
No central allocator. No political fight over which projects matter. The aggregate of individual decisions routes money to where actual usage lives. Popular infrastructure gets funded by many. Niche tooling by the few who depend on it.
Why This Compounds
Once an organisation has paid its annual contribution, the marginal cost of adopting more commons-funded software is zero. Every additional tool they move onto the scheme makes their existing contribution work harder. Procurement culture inverts. Instead of resisting every new line item, buyers actively hunt for scheme-compliant alternatives to their proprietary stack.
That’s the flywheel. The more you use, the better your deal. The more organisations join, the more software is worth building under the scheme. The more software exists under the scheme, the more compelling it is to join.
Crucially, this leaves France free to be French and Germany free to be German. Nobody has to give up sovereignty to participate. Nobody has to accept someone else’s tooling. Each country, each state, each municipality chooses what to back and how much. The scheme just ensures that when they make that choice, the money flows somewhere useful, the commons gets tended, and the downstream maintainers who are currently burning out alone get supported.
A Conversation Worth Having
None of this is fully worked out. The charter needs drafting. The licence needs real legal work — mandatory-payment licences sit outside OSI-approved open source and that fight is already live. The registry is a weekend project. The payment plumbing is the real engineering challenge, and there’s prior art worth studying — Open Collective, Germany’s Sovereign Tech Agency, the municipal co-funding efforts that already exist.
What’s new is the timing. Europe is walking away from Microsoft for the first time in a generation. The tooling, the political cover and the tolerance for disruption will never be higher than they are right now. If we miss this window we’ll spend the next decade paying for fragmented national rebuilds of the same stack, and the commons will stay as underfunded as it’s always been.
I’d rather we didn’t. If you work in public-sector IT, in an OSS foundation, in procurement, in policy — I want to hear what you’d want a scheme like this to look like, and what would stop your organisation from signing. Get in touch. I’m drafting this in the open.