Debates about digital sovereignty can be frustrating. In abstract terms, everyone agrees that we want it. Beyond that, conversations get stuck. Count yourself lucky if people around a table can even agree how to define digital sovereignty. But how to operationalize it? How to turn it into an action plan? Answers to those questions seem always just around the corner, and we never quite get there.

The problem, I think, is this: the web of digital entanglements in which we find ourselves is Byzantine. Everything hangs together: software, connectivity, hardware production chains, data storage, AI models, critical raw materials are infused into payments systems, military strategies, ubiquitous everyday tech, and so on. The day is over before you can even appreciate the complexity of the issue.

For that reason, defining or measuring digital sovereignty as a general matter is doomed to fail. It is not “a thing”, a holy grail we can find if only we dig long enough. It is a useful catchphrase for a gazillion digital vulnerabilities, asymmetries, lock-ins, and dependencies. Some of them hang together, others less so.

Building digital sovereignty therefore is not about pushing a magic button. It is about getting rid of digital unsovereignties (not really a word, but you know what I mean) one step at a time. So forget the definition. Look at the concrete problems we face, prioritize them with an eye to what matters to people, and tackle them one by one.

A digital sovereignty matrix might help. On one axis, it lists typical digital unsovereignty challenges: someone might switch off critical infrastructure. Someone might abuse citizen, government or company data. Someone might skim off profits generated here and funnel them elsewhere. Someone might lock tech users into specific tech stacks and not let them out again. Dependency on foreign entities might mean that the available tech doesn’t fit local circumstances, preferences or values (say, privacy-protecting social media). The list of challenges is longer than this, but this will do to illustrate the idea.

The other axis catalogues key societal domains in which tech is critical: education, healthcare, information provision (news), payments, navigation, and a few others. (Note that there are many domains in which tech is used heavily that do not fall into the “critical” category. GenAI to custom-tailor advertisements, for example, or algorithms that summarize football matches for news feeds. Doesn’t matter too much what tech they use…)

Put the two together, and you get a table like this:

In a next step, you can identify which of the intersections matter most, here shaded dark. For payments, the central vulnerability is disruption. For education and healthcare, it may be data abuse. For education, it may be data privacy and lock-in. Etc. This exercise generates a limited number of critical intersections on which to focus. If building digital sovereignty is about eradicating digital unsovereignties, this is where to start.

Focusing on critical intersections, rather than on “digital sovereignty” as a whole, allows us to be much more specific and gather relevant data: who dominates, say, healthcare data storage in Belgium? Who supplies critical components for navigation systems in Italy? Who has digital access to minors through software suites used in Irish high schools? Answer these questions, and you have the problem broken down into manageable pieces. You can start mapping out specific dependencies across the continent, comparing countries, universities, cities, and sectors.

Of course, overarching challenges remain. Just about everyone has an Android or iOS device in their pockets, everywhere. Dependence on a handful of US tech firms is a systemic problem, not just one of a company here or there. These issues continue to require systemic attention, for example through the Digital Markets Act.

But we set ourselves an impossible task if we think that digital unsovereignty can only be tackled in toto, or not at all. Identifying the most pressing vulnerabilities, by sector, gives us a concrete handle on what to prioritize, how to measure the problem, and where to begin fixing it.