Experts react to UK £500 million sovereign AI unit

The UK Government has launched a GBP £500 million Sovereign AI Unit to invest in artificial intelligence start-ups and infrastructure.

The fund is part of a broader push to stimulate economic growth and strengthen domestic control over strategic digital technologies. The programme centres on investment, computing resources and the creation of national data assets.

Industry executives said the move signals a UK ambition to exert greater influence over how AI systems operate and how the economic value they generate flows back into the country. The launch comes amid intensifying global competition for advanced chips, datasets and specialist talent.

Greg Hanson, Group Vice President and Head of EMEA North, Informatica, from Salesforce, said AI sovereignty extends beyond the physical location of infrastructure. He tied it to the trustworthiness of data and the impact of automated decisions on citizens and businesses.

"Sovereign AI is ultimately about control, and the ability to trust the decisions AI makes in practice.

"The UK's sovereign AI fund reflects a broader global shift toward building domestic competitive edge through investment, compute access, talent and the development of sovereign datasets.

"As AI moves toward more autonomous, agentic use cases, sovereignty will increasingly be defined by the data that powers those systems, much of which sits across public sector and enterprise environments. It's not just about who owns the models or where datasets sit, but about building a trusted data foundation that governs how models are trained, how they operate, and the decisions they make.

"That requires trusted context: data that is accurate, connected and understood across systems as those outcomes are generated. Sovereignty also relies on a skilled workforce that can challenge those outputs rather than trust them blindly.

"Without that, there is a real risk of scaling poor outcomes at speed. True AI sovereignty depends on confidence in the data and logic driving those systems, enabling the UK to harness AI at scale and strengthen its ability to attract both sovereign and foreign investment," Hanson said.

His comments reflect a growing view across the sector that national AI strategies must address the quality and governance of underlying data as much as model selection or infrastructure. Public sector datasets, in particular, are seen as central to areas including healthcare, welfare, transport and education.

Vendors and policy advisers are also focusing on how sovereign AI projects are designed and managed. They argue that common technical and regulatory foundations reduce duplication and make it easier for organisations of different sizes to take part.

Jonny Williams, Chief Digital Adviser - UK Public Sector at Red Hat, said the UK now faces a choice between repeating past fragmentation and building shared structures for AI development across government and industry.

"Sovereignty is not about isolation, it's about agency over the future, and today the Government has shown its commitment to that idea. However, to make this investment count we must avoid the fragmentation that has limited the impact of countless technology initiatives in the past.

"The priority should be establishing shared, open foundations from the start - commoditising common capabilities - so that startups, scaleups and government departments aren't rebuilding the same foundations and reinventing the wheel. That's how this investment scales, how a coherent talent and skills ecosystem develops, and how we move away from isolated pilots towards a platform the whole country benefits from.

"Success will depend less on the individual technologies this fund backs and more on the open principles that connect them, prioritising transparency, choice and control," Williams said.

Calls for shared foundations echo debates in other national AI programmes, where governments have pursued common data standards, reusable components and consistent governance frameworks. Advocates say this approach reduces waste and supports interoperability across departments and regions.

The emphasis on transparency and control also aligns with growing regulatory scrutiny of AI systems in areas such as credit scoring, recruitment and public service delivery. Regulators and civil society groups have pressed for explanations of automated decisions and for mechanisms that allow humans to challenge outcomes.

Analysts expect the Sovereign AI Unit to sit alongside existing research funding and industrial policy tools. It is likely to work with universities, established technology firms and younger start-ups focused on data infrastructure, model development and sector-specific applications.

Hanson also highlighted the role of skills in maintaining oversight of AI deployments at scale, underscoring that talent strategies will be as important as hardware procurement and direct investment.

Both Hanson and Williams highlighted the risk that poorly governed AI could undermine trust and weaken the benefits of national investment. Their remarks place data quality, open foundations and human oversight at the centre of the emerging UK debate on sovereign AI.