
☰

ORVELLAN GROUP
OPINIONS
Technology and Strategy
CLEAR STRATEGY.
SOUND DECISIONS.
PRACTICAL SOLUTIONS.
28 SEPTEMBER 2026 / Technology and Strategy
Digital Sovereignty Starts With a Credible Exit
Europe’s cloud debate is becoming more concrete. Buyers should make operational control and tested portability central to procurement, before signing the next long contract.
Illustrative photograph by Taylor Vick / Unsplash.
Opinion · 28 September 2026 · 4 min read
A company can choose where its cloud data is stored and still have very little control over its digital future. If changing provider would require rebuilding its applications, replacing essential skills and accepting prolonged disruption, the decision to remain is only partly a choice. Digital sovereignty should therefore begin with a practical question: what would it take to leave?
That question has become more pressing as buyers approach the next stage of the EU Data Act. The Commission’s explanation says the Act has applied since 12 September 2025 and that switching charges, including relevant data-egress charges, will be removed from 12 January 2027. That future date does not make migration free today, and removing a provider’s switching charge does not remove a customer’s engineering costs. The distinction should shape procurement decisions being made now. European Commission, Data Act explained
The wider European policy agenda is also turning sovereignty into a purchasing question. In June, the Commission explained the framework used in its sovereign-cloud procurement, following an April award to four providers. This is evidence that public buyers are translating the concept into selection criteria. It is not proof that a particular provider or architecture gives every customer the independence it needs. European Commission, 1 June 2026
Cloud exit planning needs an operational counterpart
The board’s concern should extend beyond the nationality on a supplier presentation. Where does operational authority sit? Who can administer the service? Which dependencies would remain if the customer changed its contract? What information and capabilities would the organisation need to restore an essential function elsewhere?
These questions do not imply that every company should run its own data centre. Specialised providers can deliver capabilities that a customer could not economically reproduce. The strategic issue is whether the resulting dependency is understood, proportionate and accepted, rather than discovered during a dispute or disruption.
The distinction is especially important when a low introductory price encourages a buyer to adopt deeply integrated services. The immediate benefit may be real. So may the later cost of replacing them. A sound procurement decision should account for both, including the people, time and parallel operation required for a move. Treating exit costs as somebody else’s future problem understates the cost of the original purchase.
A contractual right needs an operational counterpart
Portability is not established by an export button. A file can leave one system without retaining the structure, permissions or context that make it useful in another. The relevant test is whether the business function can operate at an acceptable level after the transfer.
The Commission’s February 2026 publication on interoperability illustrates why technical detail matters. It describes a study assessing specifications and gaps to support the Data Act’s interoperability framework. Its work on a Union repository and standards should not be confused with a declaration that cloud services are already interchangeable. European Commission, 23 February 2026
Buyers should turn that distinction into contract and acceptance requirements. Define the information that must be exportable, the assistance available during transition and the responsibilities for proving that transferred data remains usable. Then require a proportionate demonstration using non-sensitive test material. A successful demonstration should record its scope; moving a sample database does not prove that a complex production environment can move unchanged.
Boards also need to distinguish exit planning from an emergency recovery plan. A migration that works over several months may be commercially credible yet provide little protection against an immediate service interruption. Management should explain which problem each arrangement addresses and where residual dependence remains.
There is no universal level of independence that every workload deserves. A replaceable internal tool and a system supporting essential operations justify different spending and different evidence. The discipline is to make those choices explicit, assign responsibility and revisit them when the architecture or contract changes.
For public policy, the most useful outcome would be stronger competition around demonstrable control and practical interoperability. For companies, the immediate opportunity is simpler: use the procurement process to establish how the relationship could end while both sides still want it to begin. Digital sovereignty becomes credible when the buyer can make that decision without discovering that its business has become impossible to move.
Sources reviewed on 28 September 2026. This opinion reflects the evidence available at that date.