Digital Sovereignty vs. Digital Isolation
Earlier in this series
"Digital sovereignty" is having a moment. European ministers use it to argue for homegrown cloud providers. Regulators use it to justify data-localization laws. And critics use it as a polite word for building national intranets.
The term is doing too many jobs. Somewhere inside it, a real distinction is being blurred: the difference between a state choosing self-reliance and a state having disconnection forced upon it. I have a personal stake in keeping that distinction sharp. I have studied network constraints academically — my undergraduate thesis analyzed VPN solutions for peer-to-peer content delivery under Iranian network conditions — and I have lived on the receiving end of forced disconnection, running a consulting practice through national-scale shutdowns. Choice and force feel nothing alike from the inside. Our vocabulary should reflect that.
So here is a working taxonomy. Two questions sort almost every case.
Question one: who initiated the constraint?
Internal choice. The state itself decides to reduce a dependency: fund a domestic cloud industry, require local hosting for certain data, build national alternatives to foreign platforms. Whatever one thinks of specific policies, the state is selecting among options it holds.
External imposition. The constraint arrives from outside: sanctions that bar a country's institutions from buying software or cloud services, export controls, payment-network exclusions, or foreign vendors withdrawing service from an entire market. The state is losing options it did not choose to lose.
This axis matters because it determines reversibility and planning. A chosen constraint can be relaxed by the chooser, on the chooser's schedule. An imposed constraint cannot — which is why, as I argued in the piece that opened this series, states under imposed constraint build such different resilience architectures. They must assume the constraint is permanent.
Question two: what layer does it constrain?
Data. Localization laws requiring certain categories — financial, health, government — to be stored or processed in-country. Versions of this exist or have been debated across Russia, India, China, the EU, and elsewhere. Data can still flow; a copy must live at home.
Services. Requirements or incentives to use domestic providers for cloud, messaging, payments, identity. Europe's sovereign-cloud initiatives sit here, and so do the domestic app ecosystems that isolated states cultivate.
Infrastructure. Control over physical and logical substrate: national internet exchange points, domestic routing requirements, independent DNS arrangements, terrestrial-cable geography.
Connectivity itself. The extreme layer: throttling, blocking, or severing the link between the national network and the global internet — from filtering, through shutdowns, to a standing national intranet.
The map this produces
Cross the two questions and real-world cases sort cleanly, with the interesting cases sitting where you would not put them by instinct.
Chosen, data/services layer — sovereignty proper. EU data-residency rules, sovereign-cloud programs, most localization statutes. These are dependency-management policies inside an open network. Costly and debatable, but options remain open; a European firm under localization law can still reach the entire internet.
Chosen, connectivity layer — self-isolation. North Korea's Kwangmyong intranet is the pure case: disconnection as deliberate architecture. The state pays the full economic price of the constraint willingly. Rare, because the price is enormous.
Imposed, services layer — the sanctions cutoff. A state's banks lose access to messaging networks; its firms lose cloud accounts, license renewals, patches, app-store presence. As reported around recent sanctions regimes, whole markets have watched foreign providers withdraw within weeks. From the inside, this resembles sovereignty's mirror image: the domestic substitutes get built not from ambition but from necessity, on emergency timelines.
Imposed or hybrid, connectivity layer — the blackout. Full or partial shutdowns of national connectivity. For the population and businesses beneath it, this is pure isolation. I will keep to my standing rule and stay at the systems level here; what shutdowns do to work, family contact, and the economy gets its own piece next in this series.
The genuinely hybrid case. Iran's National Information Network is the taxonomy's stress test. It is a chosen, state-funded architecture of domestic services and infrastructure — sovereignty-shaped — built partly in response to imposed constraints, sanctions among them, and capable of operating when connectivity is constrained. Analytically it lives in several boxes at once, which is exactly why lumping "sovereignty" and "isolation" together produces such confused commentary about it.
Why the taxonomy earns its keep
Three payoffs, briefly.
It prices things honestly. Sovereignty is buying options: paying extra for a domestic alternative so it exists if needed, while keeping the global one. Isolation is losing options. Confusing them lets rhetoric hide costs — "sovereignty" sounds like strength even when a policy is quietly destroying options rather than adding them. The test I apply to any measure: after this, can actors reach more alternatives or fewer?
It explains framework blind spots. In my comparison of CIIP frameworks, I noted that OECD, EU, and AU instruments all silently assume a state inside the international system. Put precisely: they assume the top-left of this map — chosen constraints at most, at the data/services layer. They have nothing to say to states living under imposed constraints at the services or connectivity layer, which is a significant share of humanity.
It predicts architecture. Tell me a state's position on the map and I can roughly sketch its resilience strategy. Chosen-and-connected states buy redundancy. Imposed-constraint states build substitution. Connectivity-constrained states engineer graceful degradation. The map is not just descriptive; it is the independent variable in the comparative research agenda I laid out at the start of this series.
One line to keep
If you take a single sentence from this piece: sovereignty is the capacity to choose your dependencies; isolation is having the choice made for you. A policy that increases a state's choices is sovereignty even when it looks inward. A policy — or an outside act — that removes choices is isolation even when it wears sovereignty's vocabulary.
The hardest case is the bottom of the map, where the lights actually go out. That is where I go next — not as theory, but from the inside.
When the Internet Goes Dark: Governance Lessons from National-Scale Shutdowns →