UK’s sovereign AI investment raises questions over control, governance and security

The UK government’s latest investment in sovereign AI has been welcomed by technology leaders, but industry reactions suggest that funding alone will not guarantee control over the technology or the security of the systems built around it.

The Budget package includes an additional £60 million for the Sovereign AI Fund, £60 million for its R&D Procurement Scheme and £18 million for a new UK Advanced Chip Design Academy. The government says the measures will help British businesses develop AI capabilities domestically, reduce reliance on overseas technology and strengthen national security.

Sovereign AI means more than UK-based infrastructure

For Rob Demain, CEO of e2e-assure, the distinction between hosting data in the UK and maintaining genuine control over it is critical.

“The circa £140M investment in sovereign AI is welcome, but sovereignty has to mean more than where the data sits,” he said. “We’ve said for some time that data stored within our borders is no guarantee of control.”

Demain warned that foreign-owned providers could face competing obligations during a crisis, potentially leaving UK customers exposed. He argued that eligibility for priority grid connections for critical data centres should be assessed against ownership, jurisdiction and operational control, as well as technical requirements.

He also called for investment in sovereign cybersecurity, warning that the UK’s defensive capabilities could be disrupted if they depend on third-party platforms that can be withdrawn or compromised.

Cybersecurity must extend beyond national defences

For Camellia Chan, CEO and co-founder of cybersecurity company X-PHY, the investment in home-grown AI chip design, skills and security is a positive step. But she argued that national-level defences must be complemented by protection built into the hardware businesses use.

“Businesses must play their part by investing in security built into the hardware itself, providing protection against threats before they can compromise critical systems,” she said.

AI governance will determine the benefits for public services

The challenge extends beyond infrastructure and cyber defence to how AI is deployed across government. Ben Sekhon, regional VP UKI at OutSystems, said public investment will only translate into better services if departments retain oversight of the systems they introduce.

“Governance cannot be something added afterwards; it needs to be built into how AI applications are designed and deployed,” he said.

Sekhon called for transparency over the models and data being used, clear limits on what systems can do, and processes for reviewing decisions and establishing human accountability. He also warned against government departments accumulating disconnected AI tools that become difficult to monitor, update and control.

His example of Worcestershire County Council illustrates the potential benefits of a more controlled approach. There, AI verifies identification and other application documents, saving 200 hours annually while staff retain responsibility for the final decision.

Together, the reactions point to a broader test for the government’s AI ambitions. Building domestic capabilities is one part of the equation; ensuring that the resulting infrastructure, security systems and public-sector applications remain under effective control is another.

As Demain put it, the investment must become “a true British capability that is built, bought, and scaled here, rather than relying on technology owned elsewhere.”