How Data Centers Can Set the Stage for Larger Loads to Come

Data center power demand in the US is projected to reach 35 GW by 2030, potentially accounting for 7.5% of total electricity consumption. RMI argues that this surge presents an opportunity to set a precedent for future large-scale electrification (e.g., EVs, manufacturing) by prioritizing energy efficiency, demand flexibility, grid-enhancing technologies (GETs), and carbon-free energy procurement. The document highlights the risk of utilities defaulting to new gas generation and urges regulators to implement system-level 'zoning' for the grid and equitable cost allocation for upgrades.

• Data center power demand in the United States is surging, with projections suggesting it will double from 2022 levels to 35 GW by 2030. This growth is driven by AI workloads, cloud computing, and cryptocurrency mining, and is expected to account for up to 7.5 percent of total US electricity consumption, which is comparable to the energy use of 40 million homes. In 2023, there were 4.3 GW of new transactions.
• To avoid a default reliance on natural gas—with utilities reporting 16.9 GW of new gas generation planned between 2023 and 2027—data centers can employ four strategic solutions: maximizing energy efficiency, enhancing demand flexibility, optimizing existing grid infrastructure, and negotiating for carbon-free energy and storage. These actions can leverage over $100 billion in annual investment to set a precedent for future load growth from building electrification, manufacturing, and electric vehicles.
• Energy efficiency and demand flexibility are key to reducing grid pressure. Cooling typically uses 40 percent of a facility's energy, and AI-ready racks are increasing power requirements from 10–14 kW to 40–60 kW per rack; Equinix improved efficiency by 9 percent using AI for cooling. For flexibility, Verrus is using software and battery storage to reduce load during peak stress. While cryptocurrency miners respond to price signals, AI data centers may require policy mechanisms to incentivize participation in demand response.
• Data centers can accelerate decarbonization by utilizing Grid-enhancing technologies (GETs) like topology optimization and dynamic line ratings to bypass interconnection bottlenecks. Additionally, the Inflation Reduction Act's (IRA) Energy Infrastructure Reinvestment program allows for "clean repowering" by adding clean generation or storage at existing gas and coal sites. Examples of corporate leadership include Meta's green tariff with the Tennessee Valley Authority and Amazon's conversion of a Maryland coal mine into a solar farm.
• Regulators are urged to implement system-level planning—similar to a zoning system for the grid—to manage where large loads are added and minimize new infrastructure. They should ensure equitable cost allocation for grid upgrades, promote transparency in energy needs, and encourage utilities to use IRA tax incentives and GETs to limit the addition of new gas plants.

APA

RMI (2024). How Data Centers Can Set the Stage for Larger Loads to Come. https://rmi.org/resources/how-data-centers-can-set-the-stage-for-larger-loads-to-come/

Chicago

RMI. How Data Centers Can Set the Stage for Larger Loads to Come. 2024. https://rmi.org/resources/how-data-centers-can-set-the-stage-for-larger-loads-to-come/.

Wikipedia

{{cite report |author=RMI |title=How Data Centers Can Set the Stage for Larger Loads to Come |date=3 May 2024 |url=https://rmi.org/resources/how-data-centers-can-set-the-stage-for-larger-loads-to-come/ |access-date=17 August 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{rmi2024how, author = {{RMI}}, title = {{How Data Centers Can Set the Stage for Larger Loads to Come}}, institution = {RMI}, year = {2024}, month = may, url = {https://rmi.org/resources/how-data-centers-can-set-the-stage-for-larger-loads-to-come/}, urldate = {2026-08-17}, note = {Indexed by Climate Insights Directory} }