Digital Sovereignty is empty, but not boundless
A concept that means everything
Digital sovereignty is a politically charged and contested concept. In scholarship and public commentary, it is variously considered vague, underspecified, and all encompassing. In fact, it now means so many things that it should collapse under the weight of its own contradictions. For example, sovereignty as rights protection stands opposed to sovereignty as content control, both often espoused and implemented respectively in the European Union’s and China’s conceptions of sovereignty. Nor is the territorialised conception of sovereignty packaged in demands of data localisation reconcilable with conceptions that emphasise the free flow of information espoused by the United States and Big Tech corporations. In fact, contradictions reappear even when an actor holds a single conception. For example, the demand for data localisation is often expressed as protecting citizens’ data from foreign states, even as it hands the same data to the home-state. So, the question of digital sovereignty is not simply what sovereignty but also whose sovereignty.
Faced with this incoherence and fragmentation, the near-consensus position is that this is a defect: the concept is vague, empty, and must be disciplined into a single clear meaning. Sometimes, the discursive imprimatur for this position is that digital sovereignty is an “empty signifier”, invoked to criticise the term, as Julia Pohle (2026) notes, for lacking a common understanding. This is not this opinion editorial’s position. The op-ed finds that digital sovereignty’s vagueness, rather than being a defect, is productive and that this insight is made available by a discursive reading. Specifically, although recent scholarship has begun to characterise digital sovereignty as an empty signifier, emptiness is neither boundlessness nor a receptacle waiting to be “filled” by any content, as Grohmann and Barbosa (2026) think. Nor is the filling, as these authors assert, performed uniquely by corporations, as if nation-states lack agency. Rather, what a state-actor can plausibly claim as sovereignty is conditioned by political economy. This insight has policy implications for how African states and the African Union should pursue digital sovereignty.
Emptiness is not a defect
Two insights require emphasis. First, and properly understood, to call digital sovereignty an empty signifier is not a defect because emptiness is the condition of politics (Laclau, 1996, 36), i.e. what makes politics possible. Digital sovereignty is vague and fragmented because that constitutive feature is what allows it to mobilise actors – nation-states, firms, scholars, activists etc. – that are orthogonally and conflictually positioned. Only a term without a fixed content can gather actors with diametrically opposed demands. So, when scholars and policy makers lament its conceptual weakness and seek to discipline it, they are engaged in a contradiction because to insist that digital sovereignty is an empty signifier is concurrently to accept that it is constitutively vague.
Second, what conception of digital sovereignty a state can plausibly assert is conditioned – not determined – by its political economy. Put simply, the understanding that digital sovereignty has no fixed content is not a warrant that nation-states can adopt any conception of digital sovereignty. Emptiness is the structural condition through which the various demands are progressively cancelled; i.e. what distinguishes the demands from one another – their differences – becomes subordinated to a larger goal of unity against a constructed opposition (Laclau, 1996, 39). And political economy – infrastructure, capital, and legal leverage – is the variable that shapes the demands and it operates on two axes: the nature and identity of the frontier i.e. the political “them/us” divide and not the technological frontier used in digital policy. These points are borne out by empirical reality.
The evidence confirms the discursive reading
On the mobilising function of digital sovereignty as an empty signifier, the recently established China-led World Artificial Intelligence Cooperation Organisation (WAICO) supplies the empirical anchor. WAICO – an Intergovernmental Organisation (IGO) predominantly constituted by members from the Global South – is able to mobilise such differently positioned actors because, amongst other things, “its normative foundations draw… on data sovereignty”, without specifying what it is.
On the political economy dimension, consider Africa and the EU. The identity of the frontier drawn by these two regions is set by political economy. The US and China constitute the frontier because they are the only countries that control the four vectors – infrastructure, data, services, and AI models – of digital sovereignty. But this merely states the obvious. What is important is where these regions construct the frontier, i.e. its nature, even though who inhabits the frontier is the same for both. Where the frontier– influenced by political economy – is constructed, conditions the conception of digital sovereignty.
As far as Africa is concerned, the facts are indisputable: Africa holds a paltry 0.6% of global data-centre capacity; more than 80% of African data is estimated to sit offshore; and the submarine cables that connect a greater part of the continent are foreign owned. So, when Africa constructs its frontier, it settles at the lowest rung not against the US (hyperscaler) as infrastructure owner but rather against the US (hyperscaler) as data extractor. And that is why many African nations – Nigeria and South Africa inclusive – and continental frameworks gravitate towards a mix of data localisation and adequacy-conditioned data flows rather than the sovereignty conception of infrastructural autonomy. Put differently, Africa can aspire to any conception of sovereignty but the one it can purchase is tied to its political economy: legislating over the data traversing pipes and clouds it does not own.
In contrast, the EU can construct its frontier at a higher level of the rung, US/China qua infrastructure owner, and thereby frame its digital sovereignty conception as infrastructural autonomy because of its capital and industrial base. For Africa’s position, and given that it cannot construct its frontier at the ownership level, it must seek ownership partnership while resisting data extraction from the same party, thereby generating contradictory policy prescriptions. Hence, why the AU and Egypt AI strategies simultaneously want sovereignty through partnership and sovereignty against partnership.
What lessons?
First, we must concede that emptiness – as a structural condition rather than a pejorative label – is not a boundless receptacle. As the Africa and EU examples show, the conception of sovereignty is conditioned by political economy. For Africa, this suggests that a continental effort towards pooling the material resources of many nations is required in order to construct a frontier that genuinely antagonises infrastructural ownership and not the data layer. And this is where the mobilising function of digital sovereignty as an empty signifier becomes useful, just as WAICO gathered differently positioned states. Second, digital sovereignty’s conceptual vagueness is not a defect. It is the concept refracted through political economy, which is what policy makers should pay attention to because its political economy conditions the nature of its frontier including the one it can construct. Lastly, the toolbox of empty signifier – antagonism, frontier, political unity, cancellation of differences – can generate useful insights for a discursive reading of digital sovereignty.