Why digital sovereignty is a myth

Digital sovereignty is not an all-or-nothing choice. Too often, the debate swings between extremes where Europe must become fully sovereign and abandon U.S. providers, or change is too complex and we might as well all default to Microsoft, Google and Amazon. Both positions frame sovereignty as a light switch; either fully on or fully off. But sovereignty does not work that way. Framing it as a binary choice oversimplifies a deeply complex reality and leads to poor strategic decisions.

Switching from an American provider to a European one does not automatically deliver independence. Dependencies are embedded at multiple levels: data, AI models, cloud infrastructure, software ecosystems and hardware supply chains. The digital economy is a deeply interconnected system. Full sovereignty in the sense of complete independence is therefore a myth. Economically, technologically and geopolitically; complete self-reliance simply does not exist.

And focusing on that myth distracts from the real question: how much freedom to operate do we actually have?

Digital autonomy is about maintaining the ability to operate when circumstances change. It is the capacity to make meaningful choices within your digital stack, to change course when conditions change, so you are not trapped by legal, technical, or strategic dependencies beyond your influence.

That matters more than ever as AI accelerates concentration around a small number of dominant platforms. Each integration increases efficiency and deepens lock-in. The convenience is undeniable. So is the strategic risk.

Autonomy, however, is not a one-size-fits-all blueprint.

How digitally autonomous you should aim to be depends entirely on what you need to protect and what you stand to lose. A public-sector organization managing sensitive citizen data has a different risk profile than a marketing agency using generative AI for content. The task is to make conscious decisions about where dependency is acceptable, where it should be diversified and where it must be tightly controlled. Digital autonomy is therefore not a fixed destination, but a spectrum of deliberate choices.

That can take practical forms. Designing architectures that separate critical from non-critical workloads. Leveraging open-source components where appropriate. Building systems on open standards to preserve migration paths. Embedding exit clauses and portability into contracts. Maintaining enough internal expertise to understand what you are delegating, and what that delegation means five years from now.

Autonomy does not require a radical overnight overhaul: every procurement decision, every contract renewal, every architectural redesign presents a strategic moment to increase agency step by step. Not everything needs to be maximally autonomous. And not everything needs to be European or American. It's therefore not an ideological debate but just good old-fashioned risk management with a new, geopolitical take.

Organisations still argue that there are no viable alternatives to dominant hyperscalers but that argument is outdated. Today, mature European cloud providers exist, open-source ecosystems are robust and AI players beyond the dominant platforms are scaling quickly. None of them will replace a hyperscaler in a one-to-one replacement but that is not the point. Resilience doesn't come from substitution, it comes from designing modular combinations that reduce structural dependency. The problem is not market availability, but rather architectural imagination based on your risk profile, public responsibility and long-term strategy. 

That makes it more than an IT conversation: it is a governance issue, a continuity issue and a board-level risk discussion. Digital autonomy is not just one decision, it's a choice we make every day. So it isn't a plea for isolation or protectionism but for strategic resilience; the ability, as an organization and as a society, to maintain control over how we continue to develop digitally.

Tech leadership and digital infrastructure share a defining trait: both are shaped by concentrated power structures. Being one of the few women at the table during these discussions, has sharpened my awareness of how power concentrates and how rarely it redistributes itself voluntarily.

At its core, the debate on digital autonomy is about power: who controls the infrastructure, sets the standards, and determines access. When concentration is normalized, so is dependency. Diversity in leadership does not automatically solve this, but it broadens how risk and control are assessed. That plurality of perspectives ultimately strengthens strategic resilience.

Autonomy is not about standing alone. It is about standing prepared with choices intact.