In January 2016, France came within one parliamentary amendment of putting digital sovereignty on the statute book — a decade before Brussels made the phrase respectable. The amendment did not survive the month. What killed it was a wave of ridicule aimed at something nobody had proposed.
The vehicle was the loi pour une République numérique. On 6 January 2016, deputies Delphine Batho and Laurent Grandguillaume tabled amendment CL129: within months of the law’s promulgation, the government would produce a feasibility report on a future commissariat à la souveraineté numérique — an institution whose missions would serve, in cyberspace, “national sovereignty and the individual and collective rights the Republic protects.” In its initial wording the report would also examine the conditions for creating, under that body, a “sovereign operating system” and sovereign data-encryption protocols. Batho was explicit about what this meant, and said so publicly while the debate raged: an “open and democratic” operating system built on the Linux kernel and an existing distribution such as Debian or Ubuntu rather than written from nothing, serving three ends: the protection of civil liberties, national security, and economic autonomy against firms operating under foreign law.
Within three weeks, it was buried.
Gilles Babinet, who represented France’s digital interests at the European Commission, tweeted that “the idea of a sovereign OS is ridiculous and merely reveals the incompetence of its promoters.” Laurent Chemla, founder of the registrar Gandi, called it “totally surreal,” requiring “hundreds of developers and millions of euros”; the specialist site Numerama cost a new operating system at €831 million to €1.04 billion. Guillaume Poupard, head of the national cyber-security agency ANSSI, told a press conference that “if it means redeveloping an operating system from scratch, technically, it is nonsense.”The digital minister, Axelle Lemaire, said she did not “believe in the possibility of a sovereign operating system, which would be barely operative.”The Senate struck the operating-system clause; what survived, as Article 29 of the law promulgated on 7 October 2016, was a report on the mere possibility of a commissariat — the sharp edge filed off.
Read a decade later, the criticism missed its target twice over.
The first miss: a strawman. Every dismissal attacked an operating system built from scratch. Poupard’s “from nothing” described a project no one had tabled; Numerama’s billion-euro estimate priced a clean-room rewrite of a problem already solved many times over. The amendment, and Batho’s public case for it, named the opposite: a curated, hardened distribution on an existing free- software base — choosing a foundation such as Debian, configuring it to the state’s security needs, governing its updates, funding an accountable maintainer. This was already operational practice in 2016: France’s own Gendarmerie nationale ran tens of thousands of desktops on a custom Ubuntu derivative, Munich had run LiMux, India had BOSS. The critics cost a moonshot nobody wanted and presented the invoice as proof that the modest, demonstrated version was impossible.
The second miss: the wrong question. Granting good faith, the rebuttals treated “sovereign OS” as a purely technical problem, found the moonshot version absurd, and declared the sovereignty question closed along with it. Lemaire’s own formula — “technology cannot be the sole answer” was correct, and was inverted in practice: because technology is not the whole answer, the technical component was dropped altogether. That is the fallacy. Digital sovereignty needs far more than an operating system; it also needs one — technology is one indispensable part of a problem that is equally institutional, legal and industrial. The amendment understood this: it asked for an institution, a feasibility study, encryption protocols and governance, with the operating system as one component among several, never a codebase standing in for a strategy. The deeper stake was constitutional — how a state exercises authority in an a-territorial network it does not control, the subject a Conseil constitutionnel review would later treat at length.In 2016 it was met instead with a software cost estimate, as if that settled it. By April, Lemaire told the Senate she noted “with satisfaction” that the sovereign-OS question was “no longer raised,” finding it “neither opportune nor consistent with the interconnected, increasingly decentralised reality of the digital world.”That satisfaction has not aged well.
None of this was a new idea, which makes the dismissal harder to excuse. France had tried before and walked away each time: the 1966 Plan Calcul and its Délégation à l’informatique; the “sovereign cloud” of the early 2010s, whose collapse deputies themselves cited against the OS;Arnaud Montebourg’s public case for digital sovereignty as economy minister in 2014. The 2016 amendment was the first attempt to give that recurring ambition an institutional home, and it was argued away before the home was built. Each retreat left France more dependent than the last.
The deeper lesson is about who builds. The 2016 critics pictured the state writing an operating system by itself, and were right that this would be absurd. The workable model is co-production: a public anchor — sustained procurement, funding, and an institution to own the mission, precisely the commissariat the amendment proposed — resting on an existing European open-source industry and the upstream communities that maintain the common base. France had the supply side and starved it of the demand side. Mandrakesoft, later Mandriva, was a pioneering French distribution that went bankrupt partly for want of sustained public backing.Sovereignty here is a durable public commitment to a commons that European firms and communities already produce: “public money, public code,” contributions returned upstream, public software offices and practitioner networks such as France’s “Blue Hats” keeping the link alive.Without that partnership, a sovereign-OS announcement is a press release; with it, it is procurement.
The objection that finally carried — that anything ambitious belonged at European rather than “Franco-French” scale — has been vindicated by being taken at its word. The French question returned: in 2023 senators Pierre Ouzoulias and Marie-Noëlle Lienemann proposed a national OS for public administration, citing roughly 93% US share of desktop systems and close to 99% on mobile.And the European scale is now concrete. The EU OS project sketches a layered public-sector desktop — a shared European base on established open source, with national and local layers above it, deliberately built on existing distributions — as a practical complement to market-and-policy efforts such as EuroStack.What was “ridiculous” in 2016 reads, in 2026, as a roadmap.
The 2016 episode is a clean specimen of how sovereignty initiatives die: the strongest version goes unengaged while a weak misreading is ridiculed until its sponsors retreat. What has changed since is that the excuses have expired. The feasibility objection is closed by a decade of public-sector Linux running in production. The “do it at European scale” objection has been answered by Europe taking exactly that step. The cost objection has reversed — the expensive choice turned out to be the dependence, now counted in hundreds of billions of euros leaving the European economy every year. The technical and economic arguments are settled; what is left is a decision.
What must be done now is what the 2016 amendment asked for and was refused: a permanent owner for the mission. A one-off feasibility report was always on the floor. The work needs a standing institution — at EU and member-state level — mandated to fund, procure and govern a sovereign public-sector platform built on the European open-source commons, and three commitments would make it real. Procurement first: “Open Source First” and enforceable interoperability written into public tenders as pass/fail criteria, so demand actually moves. Then the ecosystem: a share of today’s proprietary-licence budgets redirected to the European firms and upstream projects that maintain the shared base — public money producing public code, sustained across budget cycles rather than announced once. Then scale: administrations pooling effort through EU OS-class platforms and shared layers, ending the waste of each ministry rediscovering the same distribution alone.
France had this within reach in 2016 and talked itself out of it. The decision now is political and European; the technical and economic questions are answered. Build the institution, write openness and interoperability into procurement, fund the commons the platform stands on — and the “sovereign OS” stops being a punchline and becomes infrastructure. The case has been made for a decade. What is missing is the nerve to act on it.
About the Author
Stéfane Fermigier is a French entrepreneur and a long-standing advocate of open source and European digital sovereignty. A co-founder and board member of CNLL (the French Open Source industry association) and of APELL, he has argued since the 2000s that Europe’s digital autonomy rests on open standards, open-source software, and the public institutions willing to sustain them. He is founder and CEO of Abilian, an Open Source software company, and a contributor to the EuroStack initiative.
Relevant Links
• Loi n° 2016-1321 du 7 October 2016 pour une République numérique (Légifrance): https://www.legifrance.gouv.fr/jorf/id/JORFTEXT000033202746
•V. Martin, “La République numérique en débat au Parlement : le projet de commissariat à la souveraineté numérique,” Nouveaux Cahiers du Conseil constitutionnel n° 57, October 2017: https://www.conseil-constitutionnel.fr/nouveaux-cahiers-du-conseil-constitutionnel
•“Un OS souverain est-il possible ?”, Le Point, 15 February 2016: https://www.lepoint.fr/high-tech-internet/un-os-souverain-est-il-possible-15-02-2016-2017965_47.php
•“Le système d’exploitation français ressurgit au Sénat,” Numerama, 2023: https://www.numerama.com/politique/1437188-le-systeme-dexploitation-francais-ressurgit-au-senat.html
•EuroStack initiative: https://www.eurostack.eu/