Argues that LibreOffice itself is competent for 90% of document work, so blaming the suite misses the point. The real lock-in is Entra ID plus a decade of integrated calendar, chat, email, and shared storage — swap out just the documents and the surrounding graph breaks, which is why users drifted back to Outlook when calendar invites stopped working.
Surfaces the Computerworld.dk report highlighting that Minister Caroline Stage Olsen's high-profile sovereignty push ended without any press release or formal announcement. The quiet reversal — staff being migrated back to Microsoft — suggests the political framing around reducing US hyperscaler dependence couldn't survive the practical reality of document fidelity and workflow degradation.
Frames the Denmark outcome as the typical ending for public-sector open-source pilots — document fidelity breaking on complex Office files, calendar and meeting workflows degrading, and a steady drip of 'close but not close enough' complaints. The pattern is cultural and operational rather than technical: users who need to get work done quietly route around the alternative stack until the project is abandoned.
Denmark's Ministry of Digital Affairs — the same ministry that spent most of 2025 publicly flirting with the idea of dumping Microsoft 365 — has quietly dropped the project. Danish IT outlet Computerworld.dk reports that the ministry's large-scale pilot of an open-source alternative, built around LibreOffice, has been shelved with no press release, no blog post, and no formal announcement. Staff who had been migrated are being moved back to the Microsoft stack.
This is the same effort that Digitalisation Minister Caroline Stage Olsen championed earlier in the year as part of a broader push for European "digital sovereignty." The framing was familiar: reduce dependence on US hyperscalers, keep public-sector data under European jurisdiction, and — not incidentally — stop writing nine-figure checks to Redmond. Germany's Schleswig-Holstein state had already announced a similar move. France's Gendarmerie Nationale famously ran on a Linux/LibreOffice stack for years. Denmark was supposed to be the next flag on the map.
Instead, the pilot ended the way most of these pilots end: not with a public failure, but with people quietly going back to Outlook because the calendar invites stopped working. Computerworld's reporting indicates the project ran into exactly the integration friction you'd expect — document fidelity breaking down on complex Office files, calendar and meeting workflows degrading, and the usual cascade of "it's close, but not close enough" complaints from users who have to actually get work done.
It would be easy to read this as a LibreOffice problem. It isn't. LibreOffice in 2026 is a genuinely competent suite — the word processor and spreadsheet are fine for 90% of what any government employee does. The problem is that the Microsoft stack isn't an office suite anymore. It's a tightly coupled graph of identity, email, calendar, chat, documents, and shared storage, all held together by Entra ID and a decade of institutional muscle memory. You can replace Word. You cannot easily replace the thing where someone drops an Excel file into a Teams channel, @mentions three people, and the shared calendar updates itself.
Every public-sector migration story of the last fifteen years has crashed on the same reef: the integration tax is paid in a currency open source doesn't mint — Exchange-compatible calendaring, Teams-grade conferencing, SharePoint-shaped document collaboration, and macro-laden spreadsheets that finance departments refuse to rewrite. Munich's LiMux project spent a decade and €14 million on this problem before reversing course in 2017. Schleswig-Holstein's own transition is still ongoing and has quietly scaled back its original timeline. France's Gendarmerie famously succeeded, but did so with a top-down military org structure and a decade of patient engineering investment — not the governance model most civilian ministries operate under.
There's also a less flattering reading, which is that "digital sovereignty" has become easier to announce than to execute. Over the past 18 months, European politicians have discovered that complaining about US cloud dependence polls well. Actually migrating is a multi-year engineering program that requires sustained political will across election cycles, dedicated staff who understand both the legacy workflows and the open-source replacements, and a tolerance for 18 months of degraded productivity while people re-learn their tools. Denmark had the headlines. It appears not to have had the budget line.
The quiet part worth saying out loud: when a government ministry can't migrate off Office, it's not primarily a procurement failure or an open-source maturity failure — it's a workflow lock-in that the industry has spent 25 years cultivating, and that nobody has a cheap exit from. The EU's regulatory pressure on hyperscalers via DMA and the AI Act is real, but regulation doesn't rewrite your organization's Excel macros for you.
It's also worth noting what the retreat *doesn't* tell us. It doesn't tell us Microsoft 365 is cheaper — the licensing math has gotten worse, not better, as Copilot gets bundled into tiers that governments can't easily opt out of. It doesn't tell us the sovereignty concerns were overblown — the CLOUD Act is still the CLOUD Act, and the EU-US Data Privacy Framework is still being litigated. It just tells us that the switching cost, right now, is higher than the political benefit.
If you're an engineer at a company (or ministry) thinking about a similar move, Denmark's unannounced retreat is a free data point. A few things worth internalizing:
Pilot the integration layer, not the apps. Nobody fails a LibreOffice migration because Writer can't bold text. They fail because the HR system exports .xlsm files, because the exec team lives in Outlook, and because the "shared drive" is actually SharePoint with permissions nobody has documented. Scope your pilot around the five workflows that touch the most people, not around a department that happens to volunteer.
Budget two years and a dedicated team, or don't start. The successful open-source office migrations in the public record — Gendarmerie Nationale, parts of the Italian military, some German states — all had multi-year horizons and dedicated internal engineering. The failures all had a procurement officer, a vendor pitch, and a 12-month target.
Separate sovereignty from suite migration. If the real goal is keeping data under EU jurisdiction, there are cheaper paths: Microsoft's EU Data Boundary, sovereign cloud offerings from OVHcloud or Deutsche Telekom's T-Systems, or confidential-compute setups that leave the apps alone. These aren't ideologically satisfying, but they ship.
Expect more of these quiet reversals, not fewer. The political fashion for digital sovereignty hasn't peaked, but the gap between announcing a migration and surviving one is becoming visible. The real test of European tech policy over the next three years won't be whether a ministry can swap Word for Writer — it'll be whether the EU can fund the unglamorous integration engineering that would make the swap actually stick. Until then, every one of these pilots will keep ending the same way: with a press release at the start, silence in the middle, and Outlook open on everyone's desktop by the time the next election rolls around.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.