Sovereignty as a Service
A phrase that refutes itself, which is exactly why we keep using it. "As a service" is the grammar of subscription. Sovereignty you subscribe to is not sovereignty. It is permission, with better branding and a monthly invoice.
We should say plainly that we did not invent the phrase, because two other groups are already using it and they mean opposite things by it.
The first group is selling it. Digital sovereignty and sovereign cloud are a real product category now, offered by Microsoft, Amazon, Google, IBM, SAP and Red Hat among others. What is on sale is a set of assurances about where your data sits and who can reach it, provided by the vendor, revocable by the vendor, and contingent on the vendor still being there and still meaning it.
The second group is criticising the first. In 2026 Rafael Grohmann and Alexandre Costa Barbosa published a paper in Media, Culture & Society called "Sovereignty-as-a-service: How big tech companies co-opt and redefine digital sovereignty". They use the phrase to name the manoeuvre: taking a word that means self-rule and selling it back as a subscription tier.
We use it a third way, as a joke that is also our entire argument. Sovereignty cannot be a service. The moment it is delivered by somebody else, on terms they set, for a fee they revise, it has become permission. The word has not been extended. It has been replaced.
Enrolled without asking
Nobody asks to be born, and yet arrival comes with obligations attached. You are a citizen of a country you did not choose, and from that follows a long list of things you are supposed to be good at being.
Philosophers have at least had the decency to argue about this one. David Hume took apart the idea of tacit consent in 1748, pointing out that a poor labourer who cannot afford to leave has not consented to anything by staying, any more than a man carried aboard a ship in his sleep has consented to the captain. That argument is nearly three hundred years old and still unresolved.
The commercial version of the same enrolment attracts no argument whatsoever. It simply happens. You buy a pair of socks, and by that act you are entered into a cohort you did not apply to join, assigned a projected value, and addressed from then on not as a person but as a likely next purchase. Both of those things have real names in the trade: the cohort is a lookalike audience, and the number attached to your head is your customer lifetime value.
You are now expected to buy the shirt. The longer version of this argument is here.
Refusal has been made expensive on purpose
The usual answer is that you can always say no. You can. Here is what saying no costs.
In a study of 300 online news outlets published at NordiCHI in 2020, every single one of the 300 offered a one-click accept. Fifteen offered a one-click refusal. Half of the sites required between ten and twelve clicks to opt out of everything.
A CHI paper from the same year scraped the five most popular consent management platforms across the top 10,000 UK websites and found that 11.8% met even minimal legal requirements. Half the sites had no reject-all button at all. And in a controlled experiment, simply removing the opt-out button from the first screen raised consent rates by 22 to 23 percentage points.
That is not friction. That is the product working as intended. The evidence is here.
Opted out is already supposed to be the default
The thing that makes this worse, rather than better, is that European law agrees with you and is simply ignored.
Article 25 of the GDPR is titled "Data protection by design and by default", and its second paragraph requires that by default only the personal data necessary for each specific purpose is processed. Recital 32 states, without ambiguity, that silence, pre-ticked boxes or inactivity do not constitute consent. Article 4 requires consent to be freely given, specific, informed and unambiguous.
Off is the legal default. Opting in was always supposed to be the affirmative act. What we have instead is a decade of interfaces engineered to make the lawful answer the expensive one, and a private citizen doing unpaid compliance work on behalf of the companies that are supposed to be doing it.
Negotiating these assumptions has become a second job, and nobody is paying for the hours.
What ownership actually restores
None of this is solved by a better privacy policy, because a policy is a promise about future behaviour by a party whose future behaviour you do not control.
Ownership is a different kind of thing. A file on your disk makes no promises, because it does not need to. There is no account, so there is nothing to suspend. No server, so nothing to subpoena. No telemetry, so nothing to leak, sell or reinterpret. The absence of a landlord is not a feature of the product. It is the whole of the product. The distinction is set out here.
This is the same point the manifesto makes about privacy: it is not a setting to be switched on, it is a property of ownership, and it left the room when ownership did.
My life
The objection underneath all of this is not really about data. It is about the assumption that a person's role is to be a good citizen and a good consumer, and that anything else is deviance to be managed.
Opting out gets treated as resistance, as though the natural state of a human being were enrolled and the burden of proof lay with the one who declines. It should be the other way round. Opting in should be the choice. Opted out should be where everyone starts.
You should not have to perform good citizenship or good consumption to be welcome at your own party.
It is your life. That is not a slogan; it is the ownership claim, and everything in this workshop follows from it.
Read further
- Enrolled without asking: consent you never gave, from Hume to the lookalike audience.
- The cost of refusal: what saying no actually costs, measured in clicks.
- Rented sovereignty: what the cloud vendors mean by the phrase, and why it cannot work.
- The Smaller Kingdom: the companion argument, about leaving fewer records behind.