- Red Hat’s survey finds AI sovereignty is a growing priority, with many UK organisations planning in case of provider exit, despite disruption risks.
- Adoption of agentic AI is high, but governance remains weak.
- Firms increasingly favour open-source approaches to boost control, transparency, and independence.
On Tuesday, 21 April, open source solutions provider Red Hat released the results of its survey on how organisations are approaching sovereign artificial intelligence (AI).
The survey, which covers 500 IT decision makers across the UK, France, Germany, Italy, and the Netherlands, indicates that AI sovereignty – the ability for organisations to have full control over their AI technologies, including data, infrastructure, and provider relationships – has become a priority for IT departments.
According to the survey, 67% of UK participants have an AI ‘exit strategy’ if their primary AI provider were to restrict access, but 43% said switching providers would have a moderate to significant impact on their business continuity.
“Our survey shows that actually executing a switch without disruption remains difficult,” said Joanna Hodgson, UK Country Manager, at Red Hat. “To close that gap, enterprises need greater control over how and where AI runs, and a consistent way to govern fast‑moving technologies like agentic AI.”
In an exclusive roundtable hosted by Red Hat, Himal Mandalia, an independent consultant working with the Digital, AI and Innovation Hub of the United Nations Development Programme (UNDP), discussed how ownership of code and AI infrastructure can influence ownership of broader systems of control.
“Open source becomes an enabler of sovereignty,” said Mandalia. He explained that changing code can change what accountability means on a daily basis, demanding a more continuous model of accountability at an organisational risk level.
Although 87% of survey respondents use agentic AI systems, only 25% have strong governance in place.
The UK’s Data (Use and Access) Act 2025 (DUAA) eases restrictions on automated decision-making by allowing it to make decisions with significant legal or other implications on individuals’ lives. It operates within certain safeguards, such as providing individuals with information regarding decisions made about them and protecting their rights to human intervention.
However, 89% of UK participants still stated that they want public policy and regulation to require open source principles to support AI sovereignty, higher than the average 77% for Europe, the Middle East, and Africa (EMEA).
Open-source principles focus on the free accessibility, modification, and distribution of software source code.
For the next three years, the respondents believe the most crucial open source benefits for building trust in AI will include greater transparency and auditability, increased tailoring for business and regulatory mandates, and more authority over how AI is built and where it operates.
AI sovereignty is also important as geopolitical shocks continue to reshape global trade, sparking increased fragmentation across regulatory norms and the regionalisation of cross-border trade. Stringent laws on data, such as the US Cloud Act and the European Union’s AI Act, offer localised mandates, leading to countries reexamining their foreign dependency for AI infrastructure.
AI sovereignty, thereby, becomes essential to consider regarding countries’ trade activities.
“Organisations are not looking for another closed, one‑size‑fits‑all stack; they want the freedom to combine different models, accelerators and clouds while staying in control,” said Hans Roth, Senior Vice President & General Manager EMEA at Red Hat.
