Policy
New AI sovereignty buyer's guide grades hyperscalers on data control
TheCUBE Research evaluated seven cloud providers across 40 offerings and found none have solved the U.S. CLOUD Act compliance challenge.
Key takeaways
- TheCUBE Research graded seven hyperscalers across 40 offerings using five sovereignty pillars: territorial, operational, technological, legal and financial control.
- No major cloud provider has solved exposure to the U.S. CLOUD Act, which allows American courts to compel data production regardless of where it is stored.
- Operational sovereignty — who responds when systems fail — proves harder to verify than territorial sovereignty, even in providers' European regions.
- Google's air-gapped offerings scored well, but legal jurisdiction risks remain when parent companies operate under U.S. law.
- AI sovereignty matters more than privacy compliance because models running on customer data directly generate revenue, making control a board-level concern.
Five pillars reveal gaps in cloud provider sovereignty claims
As artificial intelligence workloads migrate from scientific simulation to revenue-generating production systems, enterprises face a control problem that most providers have not solved. A new buyer's guide from theCUBE Research evaluates how well hyperscalers deliver sovereignty across five distinct control layers — and the findings expose significant gaps.
The assessment framework examines territorial, operational, technological, legal and financial sovereignty, according to Amit Eyal Govrin, lead analyst for sovereign AI at theCUBE Research and CEO of Agentcy Labs. Territory — where data centers physically sit — is the easiest pillar for providers to satisfy. The operational layer proves far more complex.
"You're in the AWS Frankfurt region as a German customer. Perfectly fine. Are you sovereign because you're in that region?" Govrin said during an interview at SC26. "If that region goes down, that PagerDuty may not necessarily go to somebody out of Germany. It may go somewhere else. It may go into Seattle or to Hyderabad for all you know."
The research team graded seven hyperscalers across 40 offerings, issuing 117 grades across five regions using 17 binary tests. Legal sovereignty received separate treatment because local laws make pass-fail scoring impractical. The analysis drew on more than 300 sources and hundreds of buyer interviews.
The CLOUD Act problem no provider has solved
The most significant finding applies uniformly: no hyperscaler has resolved exposure to the U.S. CLOUD Act, which allows American courts to compel U.S.-based companies to produce data stored anywhere in the world.
"No hyperscaler across the board had an answer to the U.S. CLOUD Act," Govrin said. "Now imagine you're a European entity or anywhere else outside the world, and now there's a production order from a U.S. court system that states that this hyperscaler needs to turn over your data. At that point, there is no contest. Your data becomes Uncle Sam's data."
Google's air-gapped offerings scored relatively well in the assessment, but even those configurations cannot eliminate legal jurisdiction risks when the parent company remains subject to U.S. law.
Why it matters
Unlike privacy regulations such as GDPR, sovereignty concerns directly affect revenue-generating AI systems. When models run where data sits and produce intelligence that drives business outcomes, control over that infrastructure becomes a board-level risk. The guide arrives as European regulations tighten and enterprises discover that AI success — not failure — creates the deeper problem: lock-in to providers whose legal obligations may conflict with customer interests. Companies that wait until 2027 to address these control layers will find their options constrained by systems already in production.
TheCUBE Research's John Furrier, who conducted the interview with Govrin, noted that sovereignty carries higher stakes than earlier compliance frameworks because money sits directly on the table. "When money's on the table, you know what happens," Furrier said. "Everyone fights for it or argues for it. So I think this is a tinderbox moment where one match could blow everything up."
The findings were first reported by SiliconANGLE during coverage of SC26.