CIO continue to recommend investors seize the AI opportunity, they also see growth potential across the power and resources value chain, including in utilities, infrastructure, power equipment, and storage. (UBS)
Al is accelerating data centers’ power consumption. Oracle Chairman Larry Ellison said at the launch of Stargate that the first of the project’s data centers are already under construction in Texas, and that 20 will be built at a scale of half a million square feet each. With leadership in AI an imperative for major tech companies, we believe more data centers will be built while current ones undergo capacity expansion. Data centers have always been heavy users of power and electrical equipment, but the aggregate scale and power densities are accelerating significantly due to the rise of generative AI and its use of graphics processing units (GPUs). The Electric Power Research Institute projects data centers may use up to 9% of all electricity generated in the US by 2030, up from 4% today. Globally, the International Energy Agency notes that data center electricity consumption could more than double by 2026 relative to the 2022 base of 460 TWh.
Even more electricity will be required as more advanced AI chips are developed. Computing power and server resources contribute to about 40% of the energy consumption in AI data centers. With chipmakers racing to design more advanced AI chips that are smaller and able to be more densely packed into servers for massive parallel processing capabilities, electricity consumption is likely to rise further. We estimate electrification investments in the data center end-market to reach USD 191bn annually by 2030, representing a nearly 16% annual growth rate from 2023 levels.
Grid infrastructure needs to keep up with the rising demand. Aside from data centers, the increased electrification of transport, heating, and across industries and households should also contribute to a rise in electricity consumption. And increasing electricity demand will require incremental investment in the power transmission and distribution grid. Additionally, the integration of renewable energy generation, the growth in cross-country transmission capacity, and the linkage of new power customers will require heavy investment in network infrastructure. Overall, we think transmission and distribution will experience one of the highest growth rates in electricity grid investments in the coming years.
So, while we continue to recommend investors seize the AI opportunity, we also see growth potential across the power and resources value chain, including in utilities, infrastructure, power equipment, and storage. The accelerating electrification will require at least USD 3tr of investments in annual spending by 2030, in our view.
Original report: AI value creation could amount to USD 1.16 trillion by 2027, 22 January 2025.
