Balancing cloud spend and AI deployment in Europe's digital landscape

European organisations are encountering challenges as they adopt AI, in the context of widely used cloud-first strategies. Research from Insight indicates that organisations are utilising, on average, 76% of their annual cloud capacity, with approximately 24% remaining unused. This unused capacity represents expenditure that could potentially be redirected towards infrastructure to support AI at scale.

Across the EMEA region, almost half of organisations spend up to €5 million annually on cloud services. For organisations with an average cloud spend of €3.75 million, this corresponds to approximately €901,000 in unused capacity each year, which may affect the level of investment available for AI platforms, data sovereignty measures, and infrastructure resilience.

These findings are outlined in Insight’s Digital Sovereignty Trilemma report, which describes three areas organisations are balancing as AI adoption increases:

Economic efficiency: Around a quarter of cloud capacity is not utilised, which can affect budgets available for AI development and data platforms.

Operational resilience: 47% of organisations report over-provisioning infrastructure to maintain availability.

Data sovereignty: Regulatory requirements related to data residency and governance are influencing some organisations to place workloads on sovereign or dedicated platforms, requiring trade-offs between control, cost, and performance.

AI-related demand is contributing to a reported 12% year-on-year increase in hosting costs. At the same time, a growing proportion of organisations identify digital sovereignty as a strategic priority, with this figure expected to reach 82% within three years.

Over-provisioning (47%), limited visibility (47%), and inactive resources (46%) are identified as factors associated with increased cloud expenditure and reduced flexibility. Additionally, 56% of organisations do not conduct total cost of ownership (TCO) assessments before making major workload decisions, and 41% report constraints related to legacy applications, which can affect optimisation efforts.

In response, organisations are increasingly exploring sovereignty-aware hybrid architectures. Approximately 85% are evaluating or implementing dedicated infrastructure for AI. Approaches such as improving visibility, reducing inactive resources, and aligning cloud strategies with AI and data requirements are being considered to improve efficiency.

In the UK, 78% of organisations currently consider digital sovereignty important, with this figure expected to rise to 90% within one to two years and higher over the longer term. As cloud usage expands and AI-related costs increase, organisations are reassessing workload placement to manage costs while maintaining operational requirements.