France Orders Every Ministry to Quit Windows for Linux — But Can 2.5 Million Devices Actually Make the Switch?

On April 8, 2026, at an interministerial seminar in Paris, France's digital directorate DINUM announced what amounts to the most ambitious government operating-system migration in history: every French ministry and public operator must submit a plan to exit Windows in favor of Linux workstations by autumn 2026 [1][2]. The directive covers not just desktops but collaboration tools, antivirus software, AI systems, databases, virtualization, and network equipment [3]. If carried through, it would affect an estimated 2.5 million government devices [4].

"The State can no longer simply acknowledge its dependence; it must break free," said David Amiel, Minister of Public Action and Accounts. "We must become less reliant on American tools and regain control of our digital destiny" [5].

The announcement lands in a Europe already rethinking its relationship with US tech vendors. Austria's armed forces have switched from Microsoft Office to LibreOffice. Denmark's government has committed to the same. Germany's Schleswig-Holstein is 80% through migrating 30,000 PCs to open-source software [6][7]. But France's plan dwarfs them all in scale — and in the number of things that could go wrong.

The Scope: What France Is Actually Proposing

DINUM — the Direction Interministérielle du Numérique, France's interministerial digital directorate — is coordinating the effort alongside the Directorate General for Enterprises (DGE), the National Cybersecurity Agency (ANSSI), and the State Procurement Directorate (DAE) [3]. The initiative falls under a broader digital sovereignty strategy that has already produced concrete steps: in January 2026, France ordered all state departments to replace Microsoft Teams and Zoom with Visio, a sovereign video-conferencing tool built on the open-source Jitsi framework, with full adoption mandated by 2027 [8]. The national health insurance fund CNAM has already migrated 80,000 employees to French-built alternatives including Tchap for messaging and France Transfert for file sharing [3].

The Linux migration is described as DINUM's "biggest point" [9]. Every ministry must formalize its own non-European software reduction plan covering workstations and a wide range of infrastructure. Industrial Digital Meetings are planned for June 2026 to begin coordinating implementation across agencies [3].

What France has not yet disclosed: a specific timeline for full migration, a total projected budget, or which Linux distribution it will standardize on [5][9].

The Scale Problem: 2.5 Million Devices, No Distribution Chosen

The absence of a selected distribution is both a practical and political question. The XDA Developers report notes that openSUSE — developed by the German-based SUSE company — and LibreOffice have been mentioned as "viable open-source alternatives originating within the EU" [9]. The EU OS project, a community-led initiative proposing a standardized European public-sector Linux desktop built on Fedora's immutable KDE Plasma variant (Kinoite), is actively engaging EU institutions but has no official endorsement yet [10].

Source: Various government reports

Data as of Apr 10, 2026CSV

France's government has not published an aggregate figure for its current Microsoft licensing costs. However, data from related migrations provides some scale. The French government's Visio deployment is projected to save €1 million per year for every 100,000 users in licensing fees alone [11]. Schleswig-Holstein's open-source migration is saving €15 million annually in license costs [7]. Across Europe, research by CISPE found that Microsoft surcharges for non-Azure cloud deployments cost European organizations €1.01 billion in 2022 for SQL Server alone, with an additional €560 million annually for Office 365 surcharges [12].

For 2.5 million devices, the licensing savings alone could be substantial. But licensing is only one line item in a migration budget.

The GendBuntu Precedent: France's Own Linux Success Story

France is not starting from scratch. The Gendarmerie Nationale — France's national military police force — has been running Linux on its workstations since 2008, in what stands as one of the most successful large-scale government Linux deployments anywhere [13].

The project, known as GendBuntu (a customized Ubuntu distribution), followed a gradual approach. The Gendarmerie first adopted OpenOffice.org in 2005, then Firefox and Thunderbird, before migrating the underlying operating system [13]. By June 2024, 97% of the force's workstations — approximately 103,164 machines — were running GendBuntu [14]. The total cost of ownership dropped by 40%, and annual savings reached approximately €2 million per year, with cumulative savings estimated at €50 million [14].

Critically, the Gendarmerie reported that the transition required almost no additional training for end users [14]. This is a data point that migration advocates cite frequently — but it comes with a caveat. The Gendarmerie's computing needs are relatively standardized compared to the heterogeneous software environments across France's civilian ministries, which run everything from legacy .NET applications to specialized tools built for Internet Explorer.

Source: Various government reports

Data as of Apr 10, 2026CSV

The Munich Warning: What Went Wrong With LiMux

Any honest assessment of France's plan must reckon with Munich. In 2004, the city of Munich launched LiMux, migrating 15,500 desktops from Windows to a custom Ubuntu-based Linux distribution. By 2012, the project had migrated 12,600 machines and claimed €11 million in savings [15].

Then it reversed course. In November 2017, the Munich City Council voted to return to Windows by 2020 [15].

The reasons are contested. The official justification cited user dissatisfaction and software compatibility problems. But an Accenture report commissioned by the city found that the most significant issues were organizational, not technical [15]. And a 2018 investigation by Investigate Europe, broadcast on German public television network ARD, found that a majority of city employees were actually satisfied with the operating system. Council members suggested the reversal was a personally motivated decision by Mayor Dieter Reiter [16].

The LiMux project's own post-mortem identified eight critical success factors, with political support at the top: "Without a person like Mayor Ude, and similar supporters at all levels of the administration, the whole process would have failed" [15]. When political leadership changed, the migration lost its champion.

This is the central risk for France. The current initiative has strong backing from Minister Amiel and Minister Delegate Anne Le Hénanff, who stated that "digital sovereignty is not optional — it is a strategic necessity" [3]. But French governments change. Ministerial reshuffles happen. A policy that depends on sustained political will across a multi-year rollout to dozens of ministries is vulnerable to exactly the kind of quiet erosion that killed LiMux.

The Compatibility Question: Legacy Software and Hidden Costs

The French government has not disclosed what share of its software portfolio depends on Windows-exclusive stacks. This is a significant gap in the public case for migration.

Enterprise environments typically run applications built on .NET frameworks, ActiveX controls, and Windows-specific APIs. Ministry-specific tools — procurement systems, case management software, pension administration platforms — often have deep Windows dependencies developed over decades. Each one that cannot run on Linux requires rewriting, replacement, or a compatibility layer such as Wine or a virtual machine running Windows, which partially defeats the purpose of migrating.

South Korea's experience is instructive. In 2019, the Ministry of Interior and Safety announced plans to migrate 3.3 million government PCs to Linux, budgeting 780 billion won (approximately $655 million) for the effort [17]. The ministry began pilot testing with Harmonica OS, an Ubuntu-based Korean distribution, at the Ministry of National Defense and National Police Agency [17]. But subsequent reporting has not confirmed completion of the broader migration, and the pilot highlighted the same obstacles that plagued Munich: compatibility issues with Windows-only applications and the learning curve for government workers [17].

The total-cost-of-ownership debate is far from settled. A study cited by Red Hat found that Red Hat Enterprise Linux delivered 34% annual TCO savings over Windows Server, with 29% lower infrastructure costs, 41% lower IT staffing costs, and 64% lower costs from lost user productivity [18]. But that study examined server environments, not desktop deployments — and it was commissioned by Red Hat. Independent desktop TCO analyses for government environments are rare, and the Linux Professional Institute has acknowledged that the "higher TCO" criticism of open source, while often overstated, reflects real concerns about support contracts, integration costs, and staff retraining [19].

The Enforcement Problem: How Do You Stop Ministries From Quietly Re-Buying Windows?

France's plan requires every ministry to submit a non-European software reduction plan by autumn 2026 [2]. But a plan is not a migration. The history of government open-source mandates is littered with ambitious directives that quietly collapsed at the implementation level.

Brazil's PC Conectado program promoted Linux on government machines but never achieved permanent adoption at scale [20]. South Korea's 2019 announcement has produced no confirmed large-scale rollout [17]. Italy has had various open-source policies on the books without a corresponding transformation of its government IT infrastructure [20].

The enforcement mechanism France appears to be relying on is procurement control. The State Procurement Directorate (DAE) is involved in the initiative [3], and the Visio mandate provides a template: external conferencing licenses "simply won't be renewed" after the 2027 deadline [8]. If applied to Windows, this would mean Microsoft licenses are not renewed as they expire, forcing ministries to migrate or go without.

Whether this holds depends on whether exceptions proliferate. If a ministry argues that its specialized pension software requires Windows, does it get a waiver? If enough ministries get waivers, the mandate becomes optional. France has not published an exception framework.

The Sovereignty Paradox: Is "European Linux" Actually European?

The framing of this migration as reducing dependence on "American tools" contains a tension that critics have noted. Linux itself was created by Linus Torvalds, a Finnish-American computer scientist, but the kernel's development is globally distributed. However, many of the commercial entities in the Linux ecosystem are US-based. Red Hat, the largest enterprise Linux vendor, is a subsidiary of IBM [21]. Canonical, which develops Ubuntu (the basis for GendBuntu), is a UK-based company. The Linux Foundation is headquartered in San Francisco.

SUSE, the company behind openSUSE, is German-owned and represents the strongest "European Linux" option [9]. The EU OS project explicitly aims to build a European-governed distribution [10]. But even these rely on upstream code from thousands of contributors worldwide, much of it coordinated through US-based foundations.

A 2024 paper in Science Direct examined this tension directly, finding that the EU's pursuit of "digital sovereignty" through open source creates a contradiction: the "closed" language of sovereignty sits uneasily with the "open" nature of open-source development [22]. The practical question is whether sovereignty means controlling the code you run (which open source enables, regardless of where it originates) or controlling the entire supply chain (which no single country can achieve with open-source software).

Source: OpenAlex

Data as of Jan 1, 2026CSV

Academic research on the intersection of open source, government adoption, and digital sovereignty has surged in recent years, with nearly 8,964 papers published in 2025 alone and over 46,700 total — reflecting the intensity of the policy debate across Europe [23].

The Cybersecurity Case: Both Sides

Proponents of the migration argue that running Windows exposes French government systems to risks from US intelligence access — a concern that intensified after the Snowden revelations in 2013 and has been reinforced by subsequent disclosures about US surveillance capabilities [24]. Open-source software allows government security teams to audit every line of code, something impossible with proprietary Windows binaries.

ANSSI, France's national cybersecurity agency, updated its open-source policy in February 2026, releasing all agency code under open-source licenses approved by French law [25]. This signals institutional confidence in open-source security practices.

The counter-argument is that Windows, for all its flaws, benefits from Microsoft's massive security investment — billions of dollars annually in threat research, patch development, and incident response. Linux desktop security, by contrast, depends on a patchwork of maintainers, many of them volunteers. The xz utils backdoor incident of 2024, where a sophisticated supply-chain attack targeted a core Linux compression library, demonstrated that open-source maintenance chains can be vulnerable to state-level threats [26].

Moving 2.5 million government workstations to Linux creates a large, high-value target running on distributions that may receive less security scrutiny than Windows in desktop configurations. The question is not whether Linux is inherently more or less secure than Windows, but whether the French government is prepared to fund the security infrastructure — auditing, patching, incident response — that a deployment of this scale demands.

Who Profits

A migration of this scale will create significant business for European IT service providers. French companies specializing in open-source consulting, support, and training stand to gain contracts worth potentially hundreds of millions of euros. SUSE (German), Collabora (UK-based, active in LibreOffice development), and French firms in the open-source ecosystem are likely beneficiaries [7][9].

The Schleswig-Holstein migration provides a preview: the German state allocated €9 million in 2026 to complete its open-source transition and further develop solutions for ministries [7]. Scaled to France's 2.5 million devices, the investment required for customization, training, and support contracts would be orders of magnitude larger.

What Happens Next

France has set autumn 2026 as the deadline for ministry plans — not for completed migrations. The actual rollout will take years. DINUM's coordination with ANSSI, DGE, and DAE provides institutional infrastructure that Munich lacked [3]. The GendBuntu precedent provides proof of concept within the French government itself [14]. And the broader European momentum — from Schleswig-Holstein to Austria to Denmark — means France is not acting alone [6].

But the hard questions remain unanswered. No budget has been published. No distribution has been chosen. No compatibility audit of ministry software has been disclosed. No exception framework exists. And no enforcement mechanism beyond procurement control has been articulated.

The distance between a directive and a migration is measured in years, billions of euros, and the patience of hundreds of thousands of civil servants who will need to learn new tools. France has taken the first step. Whether it finishes the journey depends on whether the political will, institutional capacity, and financial commitment survive contact with reality.