Governance – rather than cost – is derailing AI progress in UK organisations, according to Cloudera’s latest global survey, The Great AI Re-Architecture.

Cloudera found 98% of UK respondents have delayed or cancelled an AI project in the last 12 months due to data governance, compliance or regulatory issues.

At the same time, 72% say AI integration has made it more difficult to maintain effective data governance.

The findings come amid growing scrutiny of AI economics. Globally, 84% of organisations say AI workloads have increased infrastructure costs, prompting businesses to look for ways to rein in consumption and extract more value from every token – a trend sometimes described as ‘tokenmaxxing’.

Spiraling token cost is only one part of the challenge. Cloudera’s research suggests that even if organisations get AI spending under control, many still lack the data and governance foundations needed to deploy it at scale.

Almost nine in ten (89%) UK respondents believe their current data architecture requires a significant overhaul to meet future AI requirements.

Governance is also directly influencing those infrastructure decisions. Two-in-five (40%) UK respondents cite data security, governance and compliance requirements as a leading driver of changes to their AI infrastructure.

This is contributing to a rethink of where AI runs. Almost half (45%) of UK organisations have moved at least some AI workloads from the public cloud to private cloud or on-premises environments over the past year, as businesses look for greater control over costs, performance and sensitive data.

“This current era of AI is forcing organisations to rethink the foundations of their technology infrastructure,” said Sergio Gago, Chief Technology Officer at Cloudera.

Recommended reading

• AI Adoption Stumbles as Report Highlights Employee-leadership Disconnect
• UK Firms Unprepared for AI Governance Crisis
• Global AI Governance | Who’s Leading, Who’s Lagging?
• 8 in 10 Hiring Managers Prioritise AI Skills When Hiring

“Many enterprises are discovering that the architectures built for traditional analytics weren’t designed for the scale, governance, and flexibility AI demands today.

“Success will depend on building a data foundation that gives organisations the freedom to run AI wherever it makes the most sense, without compromising control or security.”

Rather than simply limiting AI consumption, the findings point to a broader infrastructure reset as organisations look to balance cost with greater control over where AI workloads and sensitive data reside.

In the UK, that shift is set to continue, with 29% planning to put greater emphasis on a hybrid-first approach over the next two years.