Energy markets rush to address AI power bottlenecks

AI’s rapid expansion is creating a surge in global electricity demand that is projected to increase by over 1 trillion kilowatt-hours annually until 2030. This growth is primarily driven by data centers, which are forecasted to contribute to almost 20% of this rise, with their power usage expected to grow by around 126 gigawatts per year until 2028, nearly equivalent to Canada’s total yearly power consumption.

Insufficient investments in electric grids have left data center developers worried about potential power shortages in 2027 and 2028, prompting capital markets investors to focus on rapidly financing energy infrastructure projects to meet the escalating requirements. Discussions at recent Morgan Stanley conferences featured talks about powering AI, highlighting the potential advantages for power suppliers and equipment companies due to data center expansions. Investors are concentrating on rising interest in off-grid solutions, resolving bottlenecks in the energy supply chain, and using credit markets to fund energy system growth.

A key emphasis at Morgan Stanley’s thematic conference in December 2025 was the necessity for a more diverse electricity grid to cater to the varied power needs of a digitized and electrified economy stimulated by AI. The CEO of an on-site electricity generation company emphasized that AI’s growth will necessitate particular types of electricity that a standard grid may struggle to provide, especially as productivity enhancements from AI could lead to a substantial increase in energy demand.

Attendees at Morgan Stanley’s Powering AI conference in November 2025 highlighted a significant transformation set to take place in the power industry over multiple decades, showcasing investment opportunities that reached a record high of $1.5 trillion in 2025. Off-grid solutions, such as natural gas, on-site renewables, microgrids, utility-scale renewables, fuel cells, battery storage, and nuclear energy, are expected to shape the future energy landscape, as the demand for AI model training exerts greater pressure on the electricity grid.

The upcoming infrastructure procurement cycle is expected to create pockets of wealth and literal power, according to Stephen Byrd, Morgan Stanley’s Global Head of Thematic Research. Natural gas is poised to satisfy approximately 20% of the world’s new power requirements, while nuclear energy and energy storage, including batteries, are attracting fresh investment interest due to their integral roles in supporting data centers and global markets.

Large-scale data centers, now reaching capacities of 1 to 4 gigawatts per site, are often impeded by regulatory, permitting, and stakeholder challenges connected to conventional grid connections, prompting a shift towards hybrid or off-grid models to ensure operational resilience. Developers aim to fortify their positions against potential energy shortages by embracing innovative power supply alternatives in a tightening global energy market environment, amplifying the focus on power equipment providers.