AI and data sovereignty is no longer a theoretical architecture debate; it is a management choice about where control sits, who bears risk, and how much dependency the enterprise is willing to accept for speed. For CIOs and transformation leaders, the key question is not whether to use third-party AI, but which workloads can tolerate external model dependencies and which require tighter control because they touch IP, regulated data, or competitive differentiators.
The practical implication is a two-speed governance model. Treat low-risk experimentation differently from business-critical deployment. That means explicit decision rights for model selection, data residency, retention, logging, and prompt/output handling, rather than leaving these issues to individual product teams. A sovereign posture may reduce exposure, but it can also narrow access to frontier capability and increase platform cost; leaders should expect that trade-off and fund it deliberately, not as an unfunded security add-on.
Portfolio discipline matters more than technology rhetoric. Next questions should include: Which AI use cases create defensible value if kept close to the enterprise data estate? Which vendors can provide contractual, technical, and operational assurances that are actually enforceable? What would a credible exit path look like if a model provider changes terms, pricing, or data-use policies? Those questions belong in governance forums, procurement reviews, and architecture boards, not only in innovation labs.
Done well, sovereignty becomes a resilience and differentiation issue. Done poorly, it becomes another fragmented platform strategy. The management test is whether the enterprise can preserve optionality while still moving quickly enough to capture AI value.
When generative AI first moved from research labs into real-world business applications, enterprises made a tacit bargain: โCapability now, control later.โ Feed your proprietary data into third-party AI models, and you will get powerful results. But your data passes through systems you do not own, under governance you do not set. The protections you rely on are only as durable as the providerโs next policy update.
Now, with generative AI established in everyday business operations and sophisticated new agentic AI systems advancing every day, companies are reevaluating the terms of that deal.
โData is really a new currency; itโs the IP for many companies,โ says Kevin Dallas, CEO of EDB, echoing a recurrent anxiety from customers. โThe big concern is, if youโre deploying an AI-infused application with a cloud-based large language model, are you losing your IP? Are you losing your competitive position?โ
That question is now fueling a movement toward reclaiming both the data and AI systems that have rapidly become part of core business infrastructure. AI and data sovereignty, which refers to breaking dependence on centralized providers and establishing genuine control over models and data estates, it is an urgent priority for many companies, says Dallas, citing internal EDB data: โ70% of global executives believe they need a sovereign data and AI platform to be successful.โ
The idea of AI sovereignty is becoming a global policy conversation. NVIDIA CEO Jensen Huang recently spoke about the need for such a shift at the World Economic Forumโs annual meeting at Davos in January 2026: โI really believe that every country should get involved to build AI infrastructure, build your own AI, take advantage of your fundamental natural resourceโwhich is your language and cultureโdevelop your AI, continue to refine it, and have your national intelligence be part of your ecosystem.โ
This report explores how enterprises are pursuing sovereignty over their models and data estates in an era of rapid AI adoption. Drawing on a survey conducted by EDB of more than 2,050 senior executives and a series of interviews with industry experts, the research confirms that the sovereignty movement on the enterprise level is already well underway.
Download the report.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Reviewโs editorial staff. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.
Related
