Digital sovereignty: the easy rhetoric and the hard supplier math

Digital sovereignty, the project of running a country or company on tech it can actually control, used to be a question of leaving. Leave the hyperscaler, leave the foreign SaaS suite, leave the US payment rails, leave the Chinese telecoms gear. Capgemini's April 2026 research suggests the leaving era is over before it began, and a quieter vocabulary has taken its place.

Fifty-nine percent of the 1,300 business and technology executives surveyed, all at organizations with more than $1B in annual revenue or budget, now regard complete digital sovereignty as unrealistic. Roughly two-thirds have redefined the goal as "resilient interdependence," Capgemini's term for managing the suppliers you cannot replace.

The shift is a receipt, not a surrender. Capgemini's Digital Sovereignty Index, drawn from 866 organizations, found 86% with significant exposure to foreign or externally controlled supply chains and only 14% reporting end-to-end visibility into the dependencies threading through their wider tech ecosystems. You cannot leave a stack whose wiring you cannot draw.

In the EU, 75% of organizations have adopted the resilient-interdependence framing; in the US, only about half have, with the remainder leaning on compliance-driven definitions. The difference tracks who actually has to choose. European buyers live inside a regulatory perimeter (GDPR, the AI Act, the Data Act, the Cloud Sovereignty Framework) that makes the conversation unavoidable. American buyers can still treat sovereignty as a procurement line item.

The same week the survey surfaced, Mistral closed a multibillion-euro round at a valuation that briefly made it one of Europe's most valuable AI labs. The headline reads like a sovereignty story; the underlying deal is the opposite. A French AI lab pitched to European public-sector buyers as the local alternative to US hyperscalers is, structurally, a managed-dependence story. The buyers get a local alternative; outside investors still get a financial claim on the stack. The story is not independence. It is the industrial policy of managed dependence, a European stack European buyers can prefer, and that outside capital can still own pieces of.

Switching cost is the obvious one: a multi-year migration, parallel run, retraining, integration breakage, and a security review window measured in quarters. Less obvious is the tier-two and tier-three problem, the long tail of middleware, observability tooling, identity providers, and vertical SaaS that nobody budgets for in a sovereignty roadmap. Capgemini's 14% visibility figure is a 14% planning horizon. You cannot exit a building whose wiring diagram you have not drawn.

What changes under resilient interdependence is the metric, not the rhetoric. A sovereignty program measured against independence fails by construction. The same program measured against concentration risk, contractual leverage, jurisdictional exposure, and the ability to swap a tier-one provider inside a quarter is tractable. Capgemini's framing implicitly asks buyers to swap targets from "no foreign suppliers" to "no single point of failure."

Governments will keep announcing sovereignty initiatives because the public demand is real and because the announcement is the deliverable. The procurement conversation underneath changes: multi-vendor architecture, contractual exit clauses, data portability, and the unglamorous work of mapping dependencies before a regulator asks.

Where the exit doors actually exist, the survey hints, is at the top of the stack. Tier-one cloud and productivity can be hedged with multi-cloud and open productivity suites. Tier-two and tier-three, observability, security tooling, vertical SaaS, identity, is where sovereignty fades, because the alternatives are thinner and the migrations are smaller but more numerous.

Resilient interdependence is a planning horizon. The next sovereignty headline a reader sees will probably be a funding round, a policy paper, or a minister at a podium. The useful question underneath is whether the buyer has drawn the map of what leaving would actually cost, and whether the contract gives them the option to do it.