89% Say Sovereignty Matters. 10% Paid For It.
BARC's Data Sovereignty 2026 is out. The headline: 89% of organizations rate data sovereignty as important. 51% say "very important." Only 2% don't care.
That's the part everyone will quote.
Here's the part they won't: 40% have invested nothing. Only 10% have a dedicated budget for the thing they just called important.
So the survey meant to prove sovereignty arrived proved the opposite. Sovereignty has arrived as a belief. It has not arrived as a budget. BARC calls execution "uneven." That's a generous word for a confessional.
The tell
Asked what they're actually doing, 10% pointed to "European Sovereign Cloud" offerings from US hyperscalers — the products with "sovereign" stamped on the box by a company incorporated under the CLOUD Act.
10%. Same as the budget number. Probably not a coincidence.
BARC says the quiet part: the modest uptake reflects "ongoing questions around verifiable control and contractual enforceability." Translation — the buyers looked at a sovereign cloud run by a foreign-controlled provider and weren't convinced. They were right.
You can't rename your way out of jurisdiction. A US provider with EU datacenters is still a US provider. The datacenter has a flag painted on it. The parent company has a subpoena waiting for it. Different flags.
90% of the market already figured this out.
Verifiable, or it didn't happen
The report keeps circling one word and never landing on it: verifiable. The hyperscaler clouds fail on it. The "implicit" strategies fail by definition — you can't verify an intention.
If your provider can read your data, sovereignty is a promise. Promises get enforced by courts, and you don't pick which government's court once your provider is foreign.
If your provider can't read it — zero-knowledge, keys never leave your hands — there's nothing to compel and nothing to leak but ciphertext. Sovereignty stops being a promise and becomes a property of the math.
One is a marketing decision. The other is a design decision. You can't bolt the second onto something built to be readable. Encryption-as-an-afterthought is the same product with a longer invoice.
The 89% are telling the truth about what they believe. The 10% are telling the truth about what they'll do. The gap between those numbers is the most honest thing in the report.
Sovereignty isn't a value you hold. It's a thing you can prove. Most of the market is still holding it.
Data Sovereignty 2026: Reality, Relevance, Roadmap
Originally published by Ross Norrie, founder of SkyeConnex, on LinkedIn.
Published May 24, 2026 · More from the SkyeConnex blog
More from the blog
Canada Needs a Sovereign Dependency Test, Because “Trust Us” Is Not an Architecture
Germany, Palantir, and the Moment “Very Impressive” Still Means “No”
Palantir just ran into a very German problem.
Read → Commentary · 3 min readSovereignty Is Not a Postal Code - CTV News
CTV ran a piece this week about Canadian companies building data sovereignty. The reporting is solid and the people…
Read → Commentary · 4 min readReassurance Is a Fine Product. It Shouldn't Cost the Same as Proof
Four Questions Every Data Sovereignty Vendor Should Answer (And Sign)
Read →