AI Investment Yields Returns as Infrastructure Confidence Stagnates

Respondents expect 63 % of existing AI investments to cover or exceed costs, with expectations rising to 89 % for future investments.

Agentic AI moves into operations. The share of organisations deploying integrated agentic AI workflows more than doubled, from 15 % to 33 %, while those still piloting isolated applications fell from 35 % to 15 %. Physical operations are also a focus: 60 % of energy-industry respondents are piloting or deploying autonomous and robotic systems for inspections and hazardous tasks.

Reliability becomes a stronger measure of value. The proportion of leaders identifying supply stability and reliability as AI’s most important contribution rose from 26 % in 2025 to 39 % in 2026. Cost-savings as the leading contribution declined from 19 % to 13 %.

Infrastructure confidence fails to keep pace. Nearly three-quarters of leaders (74 %) expect AI to increase energy consumption by 2030, but only 52 % are confident existing infrastructure can meet that demand, unchanged from 2025. Grid modernisation (67 %) and energy storage (42 %) are the top investment priorities. Confidence varies by market: 74 % in Brazil and 70 % in India consider infrastructure capable, versus 40 % in the United States and 38 % in the United Kingdom.

Brad Smith, Vice Chair and President of Microsoft, warned that progress will be measured not just by adoption but by lasting value, public trust, strong governance and sustained investment in people and skills.

Our read: The data confirms a shift from AI pilots to operational use, yet infrastructure bottlenecks threaten scalability. Companies that pair AI rollout with targeted grid upgrades and storage investments are likely to capture the reported financial upside while mitigating reliability risks.