France’s Direction Interministérielle du Numérique (DINUM) has published a formal plan to reduce dependency on non-European technology, with Linux desktop migration as a concrete pillar. The framing is deliberate: not anti-American, but pro-sovereign. The phrase used is “réduction des dépendances extra-européennes” (reducing extra-European dependencies), which is broad enough to include cloud, software licensing, and operating systems in a single policy frame.
The Hacker News thread ran predictably. Skeptics cited Munich. Optimists cited the Gendarmerie. Everyone argued about LibreOffice compatibility. The Munich comparison is understandable, but it misses how different France’s institutional position actually is in 2026 compared to what Munich was working with in 2004.
The Munich Problem Is a Political Parable, Not a Technical One
Munich began its LiMux migration in 2004, running a custom Ubuntu/Debian derivative across roughly 14,000 city government desktops. By 2013, approximately 80% of the fleet was running Linux. It was widely cited as the most significant government Linux deployment in Europe. Then in 2017, a new mayor declared it a failure, Microsoft relocated its German headquarters to Munich, and the city committed to returning to Windows. By 2021, the migration was fully reversed.
The FSFE’s post-mortem argued, with considerable documentation, that the failure was political rather than technical. The technical problems were real: legacy Windows-only applications, Active Directory integration, LibreOffice compatibility friction for external document exchange. None of these were unsolvable. All of them required sustained political will to solve rather than to circumvent by reverting to Windows. That will did not survive a change in leadership.
This is the actual lesson from Munich: government Linux migrations are governance projects that happen to have a technical component, not the other way around. The technology was never the constraint.
France Has Already Run the Proof of Concept
The French Gendarmerie Nationale began its Linux migration in 2008 under the direction of Lt. Col. Stéphane Dumond. By 2014, approximately 90,000 desktops were running Ubuntu, making it one of the largest government Linux deployments in the world. The Gendarmerie reported savings of roughly €2 million per year in software licensing costs. The migration is still running.
This matters because it is not a pilot or a planning document. It is a completed, operational deployment at a scale larger than Munich, run within the French government, under French procurement rules, with French civil servants. When DINUM references digital sovereignty as achievable, they have a concrete internal precedent to cite.
The City of Paris has similarly been running Ubuntu on a significant portion of its roughly 12,000 administrative workstations since 2017 with broadly positive results. France has already accumulated operational knowledge that Munich was still trying to develop when the political winds shifted.
La Suite Numérique Is the Other Half of the Migration
A desktop OS migration without a corresponding productivity and collaboration suite is a partial migration. This is where France is further along than most coverage acknowledges. La Suite Numérique is an already-deployed stack of open-source collaboration tools for French civil servants:
- Tchap: Encrypted instant messaging based on the Matrix protocol, with over 300,000 registered civil servant accounts. This is not a prototype. It runs on French sovereign cloud infrastructure.
- Webconf: Video conferencing based on Jitsi Meet.
- Docs: Collaborative document editing based on open-source office components, still rolling out.
- Audioconf: Audio conferencing.
None of these services depend on Microsoft, Google, or any US-operated infrastructure. DINUM has been building this stack incrementally since 2021 under its Plan d’action logiciels libres, which also created a formal Free Software Mission within the agency and launched code.gouv.fr as a portal for government open-source code.
The Linux desktop plan layers on top of infrastructure that already exists. That is a fundamentally different starting position than Munich had.
Why the CLOUD Act Made This Politically Urgent
The technical case for Linux in government has been coherent for years. The political urgency is newer, and it comes primarily from two pieces of US law.
The CLOUD Act (2018) permits US law enforcement to compel US-based cloud providers, including Microsoft and Google, to produce data stored anywhere in the world, regardless of the national laws of the country where the data resides. A French government agency storing sensitive internal communications on Microsoft Azure, even in a French data center, is operating under a legal regime where US authorities have a statutory right to request that data.
The Schrems II ruling (2020) from the CJEU invalidated the EU-US Privacy Shield, finding that US surveillance law, specifically FISA Section 702, was incompatible with EU fundamental rights protections. A successor framework, the EU-US Data Privacy Framework, was agreed in 2023 but faces ongoing legal challenges from privacy advocates including noyb.eu.
France’s ANSSI (the national cybersecurity agency) has developed the SecNumCloud qualification framework specifically to address this. SecNumCloud-qualified cloud services must be operated by entities not subject to non-EU law. OVHcloud is among the few providers that have achieved qualification. Microsoft and Google, regardless of where they locate data centers, cannot qualify because the underlying legal entities are US corporations.
The 2025-2026 geopolitical context, including US-EU trade tensions and concerns about US technology as a reliable dependency for critical government infrastructure, has given the CLOUD Act argument considerably more weight in French political discourse than it carried when the argument was primarily about abstract privacy principles.
The Challenges That Remain Real
None of this means the migration is simple. The persistent obstacles in government Linux deployments are well-documented.
Legacy line-of-business applications are the hardest problem. Government agencies run specialized software for procurement, HR, GIS, legal case management, and dozens of other domains. Many of these are Windows-only, built for Internet Explorer-era environments, and maintained by small vendors with no Linux port planned. The standard mitigations are WINE/Crossover for simpler cases, virtual machines running Windows for applications that cannot be otherwise supported, and web-based rewrites over time. All of these carry costs and timelines.
Active Directory integration is solvable with Samba 4 or migration to LDAP/Kerberos, but it requires real infrastructure work. The French government’s use of smart cards for civil servant authentication adds a specific PKI requirement that needs validation against Linux client stacks.
LibreOffice document compatibility remains a genuine friction point for any government agency exchanging complex formatted documents with the private sector. Macro-heavy spreadsheets and tracked-changes Word documents are where the compatibility breaks down most visibly. The practical mitigation is PDF-first workflows for external exchange and strict ODF templates for internal use, combined with training.
The distribution question is not yet formally settled, but DINUM’s existing guidance and the Gendarmerie’s operational deployment point toward Ubuntu LTS, likely 22.04 or 24.04, as the baseline. The French open-source community has deep Debian lineage, and Canonical offers the enterprise support contracts that government procurement requires.
What Will Determine Whether This Sticks
France has better institutional prerequisites for a sustained government Linux migration than any previous European attempt. DINUM has formal policy authority, operational deployments to learn from, a collaboration suite already in production, and geopolitical tailwinds that are not going away. Switzerland mandated LibreOffice federally in 2024. The European Parliament is considering a desktop Linux motion. The political window is wider now than at any point in the past decade.
The variable that determined Munich’s fate was not the technology. It was whether political leadership remained committed through the friction of a multi-year transition. France’s plan is framed as a long-term structural shift rather than a single migration event, and that framing, combined with the infrastructure already in place, gives it a plausible path that Munich’s 2004 effort never quite had.