Energy leaders are spending more on AI, but power supply remains a worry

Energy leaders are increasing AI investment, while many doubt the power grid can keep up. A survey by ADNOC, its investment arm XRG and Microsoft found that 89 per cent of respondents had increased AI spending over the past year, but only 52 per cent were confident existing infrastructure could meet AI-driven electricity demand.

Watch Desk analysis

What happened

The study, the third annual Powering Possible report, surveyed almost 400 senior executives in energy and finance. The National reports that 63 per cent expect returns on existing AI investments to cover or exceed their costs, while 74 per cent expect AI to increase energy consumption by 2030.

Adoption is moving beyond conventional software, too: the report says use of integrated agentic AI workflows more than doubled, from 15 per cent to 33 per cent. Six in ten energy industry respondents said their organisations were piloting or using robotics in physical operations.

Why it matters

The figures put two parts of the AI expansion in the same frame: companies are committing more money, and the infrastructure supplying electricity is not necessarily keeping pace. Confidence varies sharply by market. The report puts it at 38 per cent in the UK and 40 per cent in the US, compared with 70 per cent in India and 74 per cent in Brazil.

Respondents named grid modernisation and smart-grid technology as the leading investment priority for supporting AI growth; energy storage came next. Those are survey responses, not proof that particular upgrades will solve the problem, but they show where industry leaders say attention is going.

Our read

This is a useful snapshot of AI’s less glamorous dependency: electricity that arrives reliably, not merely a model that performs impressively in a demo. The spending figures suggest businesses expect returns; the infrastructure responses suggest the bill for making those returns possible is still an open question. The survey was commissioned by organisations with a stake in the energy and technology sectors, so its findings are best read as a report of executive views, not an independent forecast of future demand.

What to watch

• Whether the expected returns on AI investment show up in measured business results.
• Whether grid and storage spending grows alongside AI-related electricity demand.
• Whether the reported rise in agentic workflows and robotics translates into broader operational deployment.

Discussion spark: Should energy companies and governments prioritise expanding electricity supply for AI, or require AI operators to show that their workloads justify the added demand?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.