When we talk about Linux, we often think of developers, servers, distributions or personal computers.
But Linux also raises a much broader question:
who really controls our digital tools?
For an individual, choosing an operating system is often a matter of comfort, habits or available software. For a state, a school, a public administration or a local authority, the choice is more sensitive.
An operating system is not just an interface. It is a technical, economic, legal and sometimes political dependency.
This is why some institutions are interested in Linux and free software. Not only out of love for the terminal or tinkering, but because they are trying to regain some control over their tools.
Why do some institutions want to leave Windows?
Windows is extremely widespread in administrations, schools, companies and classic workstations.
This is not by chance. It is compatible with many applications. Users know it. It integrates with many professional tools. IT departments often know how to deploy, manage and maintain it.
But this dominance also creates dependency.
Dependency on a publisher. Dependency on its licenses. Dependency on its prices. Dependency on its evolution. Dependency on its ecosystem. Dependency on its commercial decisions.
For an individual, this dependency can be annoying. For a state, it can become strategic.
In April 2026, DINUM announced that it would move away from Windows in favor of Linux workstations as part of a broader effort to reduce the French state’s extra-European digital dependencies. The same statement also asks each ministry to formalize a dependency reduction plan covering, among other things, workstations, collaborative tools, AI, databases, virtualization and network equipment. (DINUM / Numérique.gouv.fr — Digital sovereignty and reduction of extra-European dependencies)
So this is not only about operating systems. It is about global digital dependencies.
Can Linux strengthen digital sovereignty?
Yes, but not by itself.
Linux can strengthen digital sovereignty because it is based on principles that are useful for a public institution:
- open code;
- possible auditing;
- adaptation to needs;
- sharing between administrations;
- independence from a single supplier;
- use of open standards;
- possibility of developing internal skills.
Free software allows an administration not only to “buy a black box”. It can understand, adapt, contribute, share, audit, publish improvements and create commons.
This is exactly the spirit of the French state’s Open Source and Digital Commons unit, which supports administrations in strengthening their use of free software, publishing and sharing source code, and building links with the open source ecosystem. (code.gouv.fr — Open source and digital commons)
But be careful: using Linux does not automatically make an institution sovereign.
If an administration installs Linux but still depends on foreign cloud services, closed business software, proprietary formats, single providers and uncontrolled collaboration tools, it has only moved part of the problem.
Digital sovereignty is not limited to the OS.
It touches the whole chain:
- operating system;
- cloud;
- data;
- business software;
- email;
- office tools;
- videoconferencing;
- cybersecurity;
- AI;
- file formats;
- internal skills;
- support;
- maintenance;
- public procurement.
Linux can be an important building block. But it is only one building block.
Why is replacing Windows difficult?
Because institutional IT is never just “a PC with a few applications”.
It is a whole ecosystem.
In an administration, you find:
- sometimes old business software;
- files and templates accumulated over years;
- office macros;
- printers, scanners, card readers;
- signature tools;
- internal procedures;
- training;
- contracts;
- user habits;
- legal constraints;
- support needs;
- invisible dependencies.
Changing OS therefore does not simply mean installing Linux.
Applications must be checked. Staff must be trained. Procedures must be adapted. Peripherals must be tested. Formats must be migrated. Support must be ensured. Exceptions must be planned. Resistance to change must be managed.
The example of Munich is often cited because it shows both sides of the issue. The city launched a large-scale Linux migration with LiMux, reaching more than 14,800 workstations and reporting savings, but it later faced difficulties around interoperability, application replacement and continuous political support. It then redirected its open source strategy toward more targeted solutions and software commons. (Interoperable Europe Portal — Munich’s Long History with Open Source in Public Administration)
The lesson is not “Linux fails”. The lesson is rather:
a successful public migration requires strategy, time, support, training and political continuity.
Without that, even a good idea can become a bad project.
Technical barriers
The first barrier is software compatibility.
Many administrations use business software designed for Windows. Some tools are old, sometimes custom-developed, sometimes linked to providers, sometimes poorly documented.
Replacing or adapting them is expensive.
The second barrier concerns files and formats.
An administration lives inside its documents: templates, forms, spreadsheets, macros, archives, exports, procedures. If everything relies on proprietary formats or very specific Microsoft Office behaviors, migration becomes heavier.
The third barrier concerns hardware.
Printers, scanners, professional cards, signature tools, specialized peripherals: everything must be tested.
The fourth barrier is integration.
A workstation is not alone. It is connected to the directory, email, network, security tools, intranet, cloud services and internal applications.
Changing the OS affects this whole chain.
Human barriers
This point is often underestimated.
For many public agents, the computer is a work tool, not a passion. They do not want to “take back control of the system”. They want to do their job without wasting time.
If the interface changes, if the software changes, if shortcuts change, if documents open differently, support is needed.
Training. Clear documentation. Responsive support. Adaptation time. Internal ambassadors. Listening to the field.
Otherwise, the migration is experienced as a punishment from above.
And then even a good system can become unpopular.
Economic barriers
Linux is sometimes presented as “free”.
That is too simple.
Licenses may cost less, yes. But a migration also costs money:
- audit;
- integration;
- training;
- support;
- software adaptation;
- change management;
- maintenance;
- documentation;
- providers;
- human time.
Free software does not remove costs. It shifts some of them.
Instead of paying only for licenses, more is invested in integration, skills, support, technical control and sometimes the local ecosystem.
It is not necessarily cheaper immediately. But it can be healthier in the long run, if properly managed.
Is the choice of an OS cultural and political?
Yes.
An operating system carries a vision.
Windows represents the dominant proprietary PC ecosystem. macOS represents closed integration controlled by Apple. Linux represents more openness, modularity, transparency and the possibility of adaptation.
For a public institution, choosing Linux can therefore mean:
- favoring open standards;
- reducing dependence on a single supplier;
- supporting local skills;
- sharing public tools;
- making some code auditable;
- strengthening transparency;
- accepting a more collaborative culture.
This is not neutral.
But we must avoid romanticism.
Digital sovereignty is not about replacing one logo with another. It is about regaining decision-making capacity.
Being able to choose. Being able to audit. Being able to migrate. Being able to negotiate. Being able to maintain. Being able to understand.
That is technological independence.
Can France really migrate to Linux?
Yes, partially. But probably not all at once, not everywhere, not instantly, and not without exceptions.
France already has important experience with free software. The National Gendarmerie, for example, became one of the emblematic cases of migration toward free tools and then toward an adapted Linux distribution, GendBuntu, with a progressive strategy that started with applications before the system itself. Articles from the time already reported a migration of tens of thousands of workstations and a logic of reducing total cost of ownership. (WIRED — French National Police Switch 37,000 Desktop PCs to Linux)
But generalizing this approach to the whole state is another scale.
The right scenario is probably not:
“everyone on Linux tomorrow”.
The right scenario looks more like this:
- identify compatible workstations;
- start with simple uses;
- move web and collaboration tools toward sovereign solutions;
- reduce office-suite dependencies;
- impose open formats;
- train public agents;
- handle business software one by one;
- keep Windows where it remains necessary;
- build lasting public expertise;
- support a support ecosystem.
It is less spectacular. But much more realistic.
Migration to Linux is not an event. It is a trajectory.
Linux as a lever, not as a flag
Linux can strengthen digital sovereignty because it allows institutions to regain control over part of the workstation.
But Linux does not solve everything.
A country can use Linux and remain dependent on foreign clouds. It can publish open source code but lack maintainers. It can adopt open formats but keep closed business software. It can announce a sovereign strategy but continue to sign large proprietary contracts.
Real sovereignty requires coherence.
Linux is therefore a lever. Not a magic symbol.
It can help regain control, provided it is integrated into a broader strategy: free software, open standards, controlled cloud, internal skills, coherent public procurement, training, interoperability, security and support for local ecosystems.
Key takeaways
Some institutions want to leave Windows because they are trying to reduce their dependence on a publisher, licenses, prices, commercial decisions and extra-European ecosystems.
Linux can strengthen digital sovereignty because it is free, auditable, adaptable and shareable.
But replacing Windows is difficult, because an administration also depends on its business software, files, procedures, users, hardware and contracts.
The barriers are not only technical. They are also human, economic, organizational and political.
The choice of an OS is cultural: it says something about control, transparency, commons and independence.
France can partially migrate to Linux, but only with a progressive, realistic and durable strategy.
Linux is not digital sovereignty by itself. But it can become an important piece of it.
