The Dutch Government Is Building Its Own Office Stack. Check Your EU Deals.

On September 25, the most discussed story on Hacker News was not a model launch or a funding round. It was a Dutch government project with a name only a ministry could love: DAWO, short for Digitaal Autonome Werkomgeving Overheid. The post pulled more than 900 points and over 500 comments in a day.

DAWO is a plan to replace Windows, Microsoft Office and the tools around them with a government owned stack built on NixOS, a Linux distribution that started at a Dutch university. Eight municipalities are already piloting it. Three state IT providers, SSC-ICT, DICTU and DUO-ICT, received the mandate to build it in July 2026. The team hopes for a stable version in 2027.

Most founders will read this as Linux news and move on. That is a mistake if any part of your pipeline sits in Europe. The reason the Dutch are doing this has little to do with licensing costs and everything to do with who holds the off switch. That same question is now showing up in procurement for companies far smaller than Microsoft.

According to reporting from Tweakers, PC Guide and Linux Stans, DAWO is not just a desktop swap. The blueprint covers four layers:

1. An operating system and workstation layer built on NixOS.
2. An office and collaboration layer, called Mijn Bureau, that blends Nextcloud with open source parts from France and Germany.
3. Cloud storage and backup.
4. A device management tool, Sextant Fleet, meant to enroll and lock down laptops at scale.

The source code is public on the Dutch government's own code platform rather than GitHub. The pilots run on old laptops the government had already written off, which removes the new hardware budget objection before anyone can raise it.

The Netherlands is also not alone. France's digital ministry is moving about 250 staff onto NixOS based machines by the end of 2026. Denmark restarted its own Microsoft free desktop project in 2025. Germany's Schleswig-Holstein has redirected part of its old Microsoft budget into open source funding.

The trigger, according to the people building DAWO, was early 2025. The United States placed sanctions on the International Criminal Court in The Hague, and the court's chief prosecutor reportedly lost access to his Microsoft services, including email.

For Dutch officials, an abstract risk became a concrete one. If a foreign government's policy could switch off one account, it could switch off a ministry. One of the project leads described the response as a kind of digital stockpile, the same instinct that makes a country keep reserves of fuel and food.

Rutger Putter, one of the core DAWO team, put the economics simply: open source "costs different." The money does not disappear. It moves from licenses to hosting, in house staff and support for the projects the stack depends on.

That framing matters for founders. Cost arguments can be undone with a discount. Sovereignty arguments cannot. You cannot discount your way out of a sanctions regime, which means this buying criterion is likely to stick.

You are not selling operating systems to ministries. Fair. But look at what DAWO actually replaces: office software, file storage, collaboration, device management. Those categories are full of startups. And the questions the Dutch asked about Microsoft are the same ones a European buyer can ask about you:

• Where is our data stored, and who can be compelled to hand it over?
• If your company is sold, or your US cloud provider changes terms, what happens to us?
• Can we get our data out in a format we can use somewhere else?
• If you disappear, can we keep running?

A year ago, many startups answered these with a paragraph in the security questionnaire. In public sector and regulated deals, that paragraph is turning into a gate. Hacker News commenters on the DAWO thread were split on whether the project will succeed, but very few argued the motive was irrational.

Note also why NixOS won inside DAWO. The team tried openSUSE and Fedora first, and both lost on ownership. openSUSE's parent has changed hands and could again. Fedora sits downstream of Red Hat, owned by IBM. The buyer did not pick the most popular option. It picked the one whose future no single foreign company controls. Your product will be judged the same way.

If you sell to European customers, especially government, health, education or finance, build a one page sovereignty answer now. It should cover five things.

1. Data residency. Which region stores customer data, backups and logs? Can a customer pick an EU only region? Be precise. "Hosted in Europe" is weaker than "primary and backup storage in Frankfurt and Amsterdam, no replication outside the EU."

2. Sub-processors. List every vendor that touches customer data, where each one is headquartered, and which ones are US companies. Buyers are starting to read this list for exposure, not just compliance.

3. Exit path. Document how a customer exports everything, in which formats, and how long it takes. If the answer is "email support," fix that first.

4. Continuity. What happens if you are acquired or shut down? Some buyers will ask for source code escrow. Others will be satisfied by open formats and a clear export. Decide which you can offer before the question arrives.

5. Deployment options. Can the product run in a customer controlled environment, even as a paid tier? You do not need to build this today. You do need a clear answer about whether it is on the roadmap.

Writing this down forces useful choices. When founders map their go-to-market in a structured plan, whether in a doc, a spreadsheet or a planning tool like Foundra, sovereignty often surfaces as a missing column: which segments will require it, and what it costs to serve them.

There is an opportunity hiding here, and it is bigger for small companies than large ones.

The DAWO project is open about needing help. Its contribution rules even allow AI coding assistants as long as a human reviews and owns what ships. Mijn Bureau stitches together components from several countries. Device management had to be built from scratch because nothing fit. Every one of those gaps is a product or a service contract for someone.

Three ways founders are likely to benefit:

Build on open formats from day one. If your product stores data in formats other tools can read, you pass the exit question without extra work. That is cheaper at seed than at Series B.

Offer an EU hosted tier early. Many infrastructure providers now make regional deployment simple. Pricing it as a premium tier tests demand before you commit.

Sell to the integrators. The Dutch are building through state IT providers and municipalities. Those organizations buy tools, training and support. A European startup that knows NixOS or Nextcloud deployment has a customer list that did not exist two years ago.

Some caution is in order. DAWO is early. Linux Stans points out that Nixpkgs, the package collection NixOS depends on, is still hosted mainly on GitHub, which Microsoft owns. France's own code for its project also sits on GitHub. Quitting a dominant vendor is slower than announcing it.

There is also Munich. The city moved thousands of desktops to Linux and later moved back to Microsoft. The difference, supporters argue, is motive: Munich's case was mostly about cost, while DAWO is about control. Whether that holds through a full rollout is an open question.

So do not rebuild your stack this quarter because of one Hacker News thread. Do assume that European buyers will ask harder questions about control, and that the ones who ask first are often the ones with the largest budgets and the longest contracts.

If you have even one European enterprise or public sector prospect, this is a reasonable month of work.

1. Week one: list every sub-processor and its headquarters country. Mark which ones hold customer data.
2. Week two: test your own export. Pick a real account, export everything, and try to load it somewhere else. Time it.
3. Week three: write the one page sovereignty answer covering residency, sub-processors, exit, continuity and deployment.
4. Week four: ask two European customers or prospects what they would need to see. Their answers will tell you whether an EU tier is a nice to have or a deal requirement.

None of this requires an engineering rewrite. It requires knowing your own answers before a procurement officer asks.

What is DAWO? DAWO is a Dutch government project to replace Windows, Microsoft Office and related tools with a sovereign stack built on NixOS. It covers the operating system, office and collaboration software, storage and device management. Eight municipalities are piloting it.

Why is the Netherlands doing this? Officials point to 2025, when US sanctions on the International Criminal Court reportedly led to its chief prosecutor losing access to Microsoft services. The goal is to avoid depending on software a foreign government can switch off.

Does this affect startups outside Europe? Yes, if you sell to European public sector, education, health or finance buyers. Expect more questions about data residency, sub-processors, exit paths and continuity.

Do I need to offer self hosting? Not necessarily. Many buyers will accept EU only hosting, clear sub-processor lists and a tested export. Self hosting is worth considering if you target governments directly.

Is open source cheaper for the buyer? Not always. As one DAWO lead put it, open source costs different. Money shifts from licenses to hosting, staff and support. That shift creates openings for service and integration businesses.

1. Dutch governments builds alternative for Microsoft based on NixOS, Hacker News discussion, September 2026
2. Dutch Government Ditches Windows for NixOS, Following France's Lead, Linux Stans
3. The Netherlands is starting to ditch Windows in favor of Linux for its government systems, PC Guide
4. Nederland maakt soeverein alternatief voor Windows en Office op basis van Linux, Tweakers
5. DAWO-NixOS source repository, code.overheid.nl (Dutch Ministry of the Interior)