What follows is my personal reading of the sovereignty question, with references to the essays I’ve written. It isn’t an anti-cloud rant: it’s the view of someone working on infrastructure that has to keep functioning over the next decade, in a geopolitical frame that isn’t helping.
Last revised: 27 May 2026.
Thesis, in one sentence
Digital sovereignty is a systemic property, not a purchase. It is measured in the number of substitutable vendors within a reasonable operational window — not in product labels or homepage flags.
Everything that follows is the reasoning behind that sentence.
How I see it
- Marketing’s “sovereign cloud” doesn’t exist. Sovereign architectures do: ones where you can exit a vendor in 6–12 months without rewriting the software. Everything else is branding.
- Europe doesn’t have a technical problem, it has a buyer problem. Public administrations buy cloud without asking what happens if the vendor changes its mind. The concept of a geopolitical kill-switch is not in today’s tender documents.
- Sovereignty and open source are not the same thing. Open source helps, but without independent operational engineering it doesn’t save you. Kubernetes is open source; your ability to staff it isn’t.
- Sovereignty isn’t only industrial — it’s already geopolitical. EU digital regulation — GDPR, DSA, AI Act, CRA, PLD — is read from the outside (San Francisco above all) as an existential attack on the American tech-capitalism model. Acknowledging that reading isn’t conceding — it’s understanding why every non-EU vendor choice is already a political decision, even when it dresses up as a technical one.
Essays on this topic
Work with me
My job here isn’t to talk you out of the American cloud — it’s to show you, with real numbers, what exiting would cost if you ever had to. It’s strategic sobriety, not a manifesto.
Who it's for
CTOs or Heads of Infrastructure at essential entities, regulated companies, or public administrations that need to defend cloud choices to a board or an authority
IT leaders at ministries, regions and healthcare about to sign (or renew) hyperscaler contracts
European SaaS founders and CTOs who want to differentiate on sovereignty without slipping into marketing
Risk committees that need to estimate exposure in non-trivial geopolitical scenarios
How I work
- Exit-readiness assessment (3–4 weeks)
I map your critical stack dependencies and estimate the cost and time to replace each dominant vendor, scenario by scenario. Output: a risk matrix with a dependency-reduction roadmap ordered by impact.
- Cloud architecture review (2–3 weeks)
I take an in-flight or planned architecture and evaluate it against the question ‘how many months to move this somewhere else?’. Useful before multi-year commitments.
- Strategic decision coaching (ongoing)
A couple of calls per month for the heaviest cloud decisions: onboarding a new vendor, negotiating an exit, public positioning on sovereignty.
Engagement FAQ
- Are you against AWS, Azure, GCP?
No. I’m against using them unconsciously. In most cases a hyperscaler is the right choice — as long as you know what you’re buying and what losing it would cost.
- Do you only work with the public sector?
No. But the public sector, essential entities and regulated industries are where the demand for sovereignty is most concrete.
- How long does a typical engagement last?
Three to six weeks for an assessment or review, ongoing for strategic coaching.
- Do you also do migration?
No. Independent sovereignty advisory works precisely because it doesn’t sell capacity. If execution is needed, I support it — I don’t perform it.
Email me at hello@margiovanni.it with a couple of lines of context. I reply within a few business days with a concrete proposal, or a polite no if it's not my scope.
Questions & answers
What do you mean by 'sovereignty'?
The capacity to keep delivering a service — public or private — when a non-EU vendor changes its pricing, its terms, or is compelled by its own country’s law to do something inconvenient for us. It isn’t autarky: it’s having negotiating leverage that we largely don’t have today.
Is the American cloud the problem?
The problem is concentration. If 70% of a country’s critical workloads run on three hyperscalers, your digital policy is effectively co-decided in Redmond, Mountain View, and Seattle. Regardless of technical quality — which is excellent.