The idea of a fully sovereign information system is both appealing and thought-provoking. Yet, in a globalized digital ecosystem, the real challenge is no longer eliminating all dependencies, but deciding which ones should be managed as a priority.
Whenever a technological crisis emerges, digital sovereignty returns to the forefront of public debate. Dependence on American hyperscalers, the consequences of Broadcom's acquisition of VMware, the rise of artificial intelligence, and growing geopolitical tensions all remind us how heavily our digital infrastructures rely on globally distributed value chains. In this context, the temptation to seek a fully sovereign information system is understandable.
In Search of an Impossible Sovereignty?
Such an ambition is understandable in terms of reasoning. However, today it is largely out of reach, as was recently pointed out during a debate between experts. Processors, GPUs, operating systems and certain essential components remain dominated by a few non-European players. According to several converging estimates, an overwhelming proportion of the GPUs used for artificial intelligence are still produced by Nvidia today, illustrating the level of Europe's dependence on this technological layer. This is, moreover, what the Draghi report on the future of European competitiveness indicated a few months ago. Should we conclude that digital sovereignty is an illusion? Certainly not. But we undoubtedly need to change our perspective.
Moving Beyond a Binary Vision of Sovereignty
Digital sovereignty cannot simply be declared; it must be built gradually. A sovereign organization is not one that depends on no one—it is one that understands its dependencies, prioritizes them, and retains the ability to evolve them over time.
Rather than opposing "fully sovereign" infrastructures to those that are not, organizations should adopt a layered approach. Not all dependencies carry the same strategic importance. Some can be reduced quickly, while others will require several more years of industrial development.
This approach first requires a precise understanding of the information system itself. Which applications are truly critical? Where is sensitive data located? Which suppliers represent the most significant dependencies? Mapping these elements has become an essential prerequisite for any credible sovereignty strategy.
Some technological layers already offer strong European alternatives. Cloud services, virtualization solutions, OpenStack platforms, and certain collaboration suites are progressively helping organizations diversify their dependencies. Other areas remain far more complex. Semiconductors, AI accelerators, and operating systems are still largely dominated by a handful of global players.
The challenge, therefore, is not to eliminate every dependency but to prioritize, understand, and reduce those that pose the greatest risks to business continuity.
Reversibility Becomes a Governance Principle
This shift is also changing the way digital architectures are designed. For many years, organizations focused primarily on performance and cost. Today, the ability to preserve freedom of choice has become just as important.
The recent VMware experience demonstrated how a technological dependency can quickly become an economic dependency when contractual conditions change abruptly. Designing open architectures, favoring standards over proprietary technologies, anticipating migration scenarios, and planning exit conditions from suppliers from the outset are now essential principles.
Reversibility is no longer merely a contractual clause—it has become a cornerstone of resilience.
It is also worth emphasizing that this transformation is primarily a human challenge. The technologies often already exist. What is lacking most are the skills required to deploy, operate, and evolve them. Training teams, developing expertise in open technologies, and supporting organizational change have become just as crucial as technological investments themselves.
From this perspective, sovereignty cannot be improvised; it must be built through a structured and methodical approach. Before making any technological decision, organizations should audit their digital dependencies, map their critical applications and data, and define a sovereignty roadmap aligned with their business priorities.
Only after this assessment does it become meaningful to implement technical transformations, prioritizing the most sensitive assets and organizing their gradual evolution within a realistic framework of control and risk management.
Sovereignty Built on Trust
Ultimately, digital sovereignty is not determined by the location of data hosting or by the nationality of suppliers. It is defined by an organization's ability to maintain control over its choices over time.
Achieving this requires dependency governance, modular architectures, genuine reversibility, and trusted relationships between organizations and their partners. This aspect is essential. There is no universal formula: every information system has its own history, constraints, and exposure levels.
At a time when geopolitical tensions are intensifying and digital infrastructures are increasingly becoming strategic assets, sovereignty is no longer a promise of technological purity. It is a pragmatic, long-term approach to risk management.
In reality, true sovereignty is probably not about eliminating every dependency. It is about understanding them, choosing them whenever possible, and never being forced to endure them.
