What is AI sovereignty and why are companies chasing after it?

Companies and countries increasingly want to call the shots when it comes to AI. The question is: Can they realistically do so?

Research firm IDC predicts 75% of G2000 companies (i.e., Forbes’ ranking of the top 2,000 global public companies) will “prioritize pursuits of AI sovereignty, using a blend of non-public hosting, open technologies, and regional partners to support mission-critical AI uses” by 2027.

What is AI sovereignty? Juan N. Pava, a research fellow at the Stanford Institute for Human-Centered Artificial Intelligence, told IT Brew the concept of AI sovereignty is not “very well-defined.” At a high level, though, it means companies and countries having some level of autonomy over AI tech stacks.

“It’s really about a country’s ability to choose and shape its dependencies on AI systems,” Pava said. “And this is inclusive of across the entire AI tech stack, so, compute talent, models, applications, etcetera, and ensuring that it’s aligned with strategic goals for a country or company.”

Why AI sovereignty is piquing interest. Pava said countries pursue AI sovereignty with four main objectives in mind: cultural autonomy, national security, economic competitiveness, and regulatory oversight.

For companies, the goals are slightly different. Hanah-Marie Darley, co-founder and CAIO at Geordie AI, told IT Brew some businesses are interested in AI sovereignty to reduce vendor dependency.

“One of the things that [companies] might be thinking about, for example, is if I partner with…insert AI company here, how locked in am I going to be to their terms and conditions as my business needs change?” Darley said. “And also, how interdependent are those systems going to get for the rest of my business?”

Businesses might also be interested in more control over their AI tech stack to keep their crown jewels protected. A June 2025 IDC survey found almost three in 10 organizations (29%) are concerned about data or IP loss from improper AI use.

“It’s not just regulatory compliance, it’s not professional governments, it’s also a competitive standpoint,” Dave Pearson, IDC group VP and global lead of core infrastructure, said at the 2026 IDC Directions conference after presenting the survey results.

How realistic is AI sovereignty? Countries and companies seeking full control over AI tech stacks may be left disappointed, according to Pava and Darley.

“For most countries, even actually for all countries, it’s neither realistic nor necessary,” Pava said. “I think controlling the full staff from chips to models and applications, it requires enormous capital, talent, industrial capacity that [only a] handful of countries, pretty much two countries, possess.”

Darley said businesses will likely need to achieve AI sovereignty initiatives by partnering with other companies or countries as opposed to standing up their own models.

“Even if you can guarantee the data storage, you won’t guarantee the functionality, and so you probably won’t keep pace with frontier models that have billions in investment if you’re trying to build them yourself,” she said.

What’s an IT pro to do? Darley said it is important for companies to get a better understanding of where AI lives in their supply chain as they embark on AI sovereignty objectives: “If you know better, you can plan better, even if you can’t do better right now.”