Energy Leaders Boost AI Spending, but Doubt Grid Capacity

What happened

Energy companies are spending more on artificial intelligence and are starting to see returns, but many leaders doubt that existing power infrastructure can keep up. A survey led by Adnoc found that about 89 per cent of energy leaders increased their AI spending over the past year.

The findings come from the Powering Possible report, the third annual study by Adnoc, its international investment arm XRG and Microsoft. It was published on Thursday and drew on responses from almost 400 senior executives in the energy and finance sectors, according to reporting by The National.

Confidence in the grid is mixed. Around 52 per cent of respondents believe existing infrastructure can meet the electricity demand AI will create, the same proportion recorded in 2025. Meanwhile, 74 per cent expect AI to push energy consumption higher by 2030, and 63 per cent expect returns on existing AI investments to cover or exceed their costs.

Why it matters

The study suggests energy operators are moving AI from experimentation towards measurable return. Investment is now a central business decision, and it sits alongside questions about whether the physical systems supporting it are ready. For leaders running large, asset-heavy operations, AI planning increasingly has to account for power capacity and resilience as well as software capability.

The flat confidence figure is the detail to watch. Expected AI-driven demand is rising, yet confidence in existing infrastructure has not moved since 2025. That gap points to a constraint that is as organisational and physical as it is technological. For experience and transformation leaders, digital services built on AI need to be tested against real-world capacity and operational resilience, not only against product roadmaps.

By the numbers

• 89 per cent of energy leaders increased AI spending in the past year.
• 63 per cent expect returns on existing AI investments to cover or exceed their costs.
• 52 per cent are confident existing infrastructure can meet AI-driven electricity demand, unchanged from 2025.
• 74 per cent expect AI to increase energy consumption by 2030.
• Almost 400 senior executives from the energy and finance sectors took part in the survey.

The Renascence take

Most readers will focus on the 89 per cent spending figure and overlook the tension beneath it. Investment is accelerating faster than confidence that the underlying power can support it.

The real signal here is not that energy leaders are spending on AI; it is that roughly half of them doubt the grid can carry the result. Service design teaches that a promise is only as credible as the weakest link that delivers it, and power capacity has now become that link. Operators should stress-test AI-enabled services against peak demand and failure scenarios before scaling them, and treat infrastructure readiness as part of the customer experience. The advantage will go to whoever keeps the service running while the models grow.