UK firms tie AI adoption to sovereignty & energy use

UK businesses are more likely to adopt artificial intelligence platforms when providers can meet data sovereignty requirements and demonstrate progress on energy efficiency, according to new research from Argyll Data Development.

The survey found that 64% of respondents would be more likely to adopt an AI platform if it met sovereignty requirements. The results reinforce a growing view among technology buyers that infrastructure choices are driven by more than performance alone.

Respondents also questioned whether the UK has enough domestic computing capacity for wider AI deployment. Some 38.9% said the country lacks sufficient domestic compute capacity, and a further 32.8% said the shortfall is already creating risk.

Together, these concerns point to a market in which procurement decisions link technical architecture to compliance, operational control and resilience. They also reflect the UK's place in a wider global competition for AI infrastructure, where access to chips, data centre power and specialised facilities has become a board-level issue for many organisations.

Sovereignty focus

Beyond adoption intent, the research points to a broader emphasis on where critical AI infrastructure sits and who controls it. More than seven in ten respondents said it is important that critical AI infrastructure is hosted within their own country or region.

Regulated industries have long managed restrictions on data residency and cross-border processing. The survey suggests these considerations are now influencing AI purchasing across a wider range of sectors. It also indicates that organisations increasingly view sovereignty as a risk-management issue, alongside vendor concentration and geopolitical exposure.

Argyll Data Development described the findings as evidence that sovereignty is becoming a commercial factor as well as a compliance requirement. It said businesses see practical benefits in local hosting and domestic capacity, including clearer oversight of data handling and more predictable alignment with UK regulatory expectations.

Debates around sovereign AI have expanded over the past year as organisations have moved from experimentation to deployment. Buyers are paying closer attention to where model training happens, where inference workloads run and how data moves across supply chains. Many also want assurance on incident response, auditability and access controls when AI systems sit inside critical workflows.

For UK-based organisations, the issue is sharpened by a perceived gap in domestic compute. If local capacity is constrained, companies may face a trade-off between speed of deployment and the degree of control they can maintain over data and operations.

Energy scrutiny

The research also points to energy as a growing part of AI decision-making. Some 51.2% of organisations said energy efficiency is important when considering AI strategies.

Respondents indicated that external pressure is already shaping how companies think about power use. More than half said customers or stakeholders expect organisations to address the energy usage of AI systems; within that group, 21.0% said this is an expectation from many stakeholders.

Another 28.6% said they believe energy concerns will become an issue in the future. This suggests businesses expect closer scrutiny of AI-related emissions and power demand, even when systems are procured through third parties.

Energy efficiency is also closely linked to cost. As AI use shifts from pilots to wider roll-outs, organisations must account for ongoing operational expenditure, including power and cooling in data centres. For many IT and finance teams, those factors sit alongside performance metrics and service-level commitments.

Infrastructure plans

Argyll Data Development is positioning itself to meet emerging demand for UK-hosted AI infrastructure, with a focus on local energy generation. It is developing large-scale UK infrastructure using renewable power and designed for sovereign operation.

It highlighted the 184-acre Killellan AI Growth Zone in Argyll, Scotland. The project includes on-site wind, wave and solar power alongside data centre engineering. Argyll Data Development said the development aims to deliver net-zero digital capacity and create skilled regional employment.

The survey arrives as government and industry debate how to expand compute in the UK without placing disproportionate strain on the power grid. Developers have proposed approaches ranging from building near existing generation capacity to investing in new grid connections and local renewables. The results suggest enterprise buyers increasingly consider energy sourcing and operational efficiency when selecting suppliers.

Peter Griffiths, Chairman of Argyll Data Development, linked sovereignty and energy efficiency directly to adoption rates.

"The message from UK businesses is clear: AI adoption will scale fastest when it is sovereign, efficient and sustainable," said Peter Griffiths, Chairman of Argyll Data Development.

"Sovereign, energy-efficient AI isn't about slowing innovation - it's about making AI viable at national and enterprise scale. The organisations that get this right will be the ones best positioned to turn AI ambition into long-term value," Griffiths said.