Europe Has the Cards. It Has Always Had the Cards.

His latest piece is correct on every count. ASML is irreplaceable. The internal market is enormous. OVHcloud, Scaleway, T-Systems and Orange all exist and would like your business. GDPR is the most successfully exported piece of legislation on the continent. Europe has the means.

What Europe doesn't have is the diagnosis.

Sovereignty is not a procurement category

The reflex is to treat sovereignty as a shopping problem. We need a European cloud. A European chip. A European search engine. A European model. A European toaster. After twenty years the score is roughly hyperscalers 11, Gaia-X nil.

This isn't bad procurement. It's a category error. Sovereignty is a property of an arrangement, not a SKU on a framework agreement. You can't tender for it. You especially can't tender for it from the country whose technology you're trying to stop depending on.

The "sovereign cloud" is a costume, not a discipline

The compromise position is the European region of an American hyperscaler. AWS Frankfurt. Azure Paris. Microsoft has actually shipped a product called the Cloud for Sovereignty, and Microsoft sells it. I will leave that one there.

The mechanics: American company, American software, hardware in Europe, data readable to the operator, operator subject to the CLOUD Act. The CLOUD Act compels disclosure on lawful request, regardless of where the bits sit, and regardless of whether the customer is told.

Calling that arrangement sovereign is not a lie exactly. It is the same kind of true as describing a rented suit as "owning a wardrobe."

The trick that has been sitting on the desk

The primitives that make the operator's jurisdiction strategically irrelevant are not new and are not European. Zero-knowledge encryption. Erasure coding. Shard distribution across non-colluding providers. Apply them properly and the question of whose cloud you're on stops being interesting. The provider can be subpoenaed; what they hand over is ciphertext fragments under keys they have never seen.

This is sovereignty as a property of architecture, not procurement. It is the thing Breton keeps almost saying.

It is also what makes the reversibility clauses he praises actually mean something. Reversibility on paper is a clause no enterprise has ever successfully invoked, because invoking it means re-platforming a multi-petabyte estate over eighteen months without going bankrupt. Reversibility that works looks like a redistribution of shards: a Tuesday afternoon, an inconvenience, not a corporate near-death experience.

That isn't policy. That is math.

The unhelpful conclusion

Europe does not need a five-year industrial strategy to win sovereignty. It needs to stop describing sovereignty as a procurement problem and start describing it as a math problem. The hyperscalers are not the threat. The hyperscalers being able to read your data is the threat. Solve the second one and the first one collapses into vendor selection.

The trump card is encryption that cannot be compelled, distributed across providers that cannot collude, under keys that never leave the jurisdiction that issued them. It has been sitting in the deck the entire time.

Brussels keeps drawing another card.

https://www.linkedin.com/pulse/digital-sovereignty-europe-has-cards-shape-balance-power-breton-8ayye/

Originally published by Ross Norrie, founder of SkyeConnex, on LinkedIn.

Published May 5, 2026 · More from the SkyeConnex blog