Tag: Load Growth

Blog Post: The Implications of AI-Driven Load Growth, as published by Environment+Energy Leader

And what it means for clean energy buyers and sellers

Authored by Jennifer Newman, PhD, VP, Research, REsurety

Load growth is dominating headlines and decision-making across the American energy industry in 2025. The artificial intelligence (AI) arms race between the world’s largest and wealthiest tech companies is driving hundreds of billions of dollars into the construction of new data centers1 and high-performance computer equipment manufacturing facilities2. Data centers and high-tech manufacturing represent the largest segments of demand growth3 in most wholesale power markets (but are by no means the only driver of growth4) and projected near-term load growth has increased substantially year-over-year5.

Power prices and price volatility are likely to rise in the short-term in response to this developing surge in demand. Any firm looking to transact in electricity or build new generation resources will encounter heightened uncertainty and risks. Typical strategies for clean energy buyers in particular could be exposed to significantly more upside or downside than was originally anticipated. However, barring wider changes in economic conditions or geopolitics, we contend that these increases will not immediately make or break existing procurement strategies and hedges. While it may feel as though power market conditions are changing rapidly, there are reasons for skepticism and caution. This article highlights three factors that we think will modulate the impact of data center-driven power market evolution: large load interconnection rates, resource adequacy preservation, and uncertainty in data center efficiency gains.

The first market dynamic to consider is the rate of large load interconnection. Just as new generators need to apply to ISOs and RTOs to join the grid, large new loads also need to apply for interconnection. Over the last decade, generation interconnection queues across the United States have grown ten-fold to over 2,600 GW, driven primarily by new wind, solar, and storage projects. These generation resources are not being reviewed and approved for interconnection to the grid nearly as quickly as they have been added to the queue, resulting in years-long backups and delays in most regions. As new data centers apply for grid interconnection in unprecedented volumes, it is likely that queue times for these projects will also grow substantially unless market reforms are implemented and supply chains can keep pace. Some data centers may elect to develop new generation capacity on-site to help ease the interconnection burden or, in extreme circumstances, operate off-grid as has been considered by some prospective green hydrogen producers. It is also likely that some proposed data centers will not make it to market at all due to a range of grid and non-grid-related factors, a pattern that is well established in the generation queues. Data center interconnection attrition will therefore temper the growth of electricity demand and, all else being equal, power prices.

The second market dynamic to focus on is the relationship between overall power supply and demand, otherwise known as resource adequacy8. Grid operators are tasked with maintaining resource adequacy and a healthy surplus of generation capacity to cover for generators that are offline for maintenance or other reasons. Existing generation resources, particularly older, ‘dispatchable’ coal and gas units, retire for economic or regulatory reasons every year and need to be replaced with new generation to maintain resource adequacy. Load growth and more frequent extreme weather conditions are two of the many other factors that threaten to further erode resource adequacy. In response, most U.S. power markets (with ERCOT as the primary exception) operate capacity markets, which both pay for generation capacity to remain online and send price signals to generation operators and developers when more capacity is needed in the coming years. The recent record-high capacity prices recorded in the 2025/2026 PJM Base Residual Auction results9 are a stark example of this signal. While there is plenty of new generation capacity that wants to connect to the grid, very little of that capacity is dispatchable.

Grid operators discount the true capacity of resources10 like wind, solar, and battery storage to better account for their contribution to power supplies during high demand periods. Data centers in particular reinforce this pressure due to their typically steady 24/7 demand profiles. In a rapid load growth environment, older dispatchable resources that might otherwise retire would be incentivized to stay online for an extended period to support grid reliability11 and capture higher power prices. While the pace of load growth is likely to outstrip generation additions and extensions of existing resources (including mothballed nuclear power plants12 in some extreme cases), grid operators will not sacrifice resource adequacy for the sake of accommodating every data center that wishes to interconnect. As long as resource adequacy is maintained in the coming years, wholesale power prices in a given market are unlikely to rise by leaps and bounds.

Another factor to consider in data center load growth is energy efficiency. New AI models like DeepSeek have grabbed headlines in recent weeks with their apparently substantial energy efficiency gains compared to incumbent AI models13, raising questions about the relationship between AI energy efficiency gains compared to AI service demand growth14. Although no one knows exactly how much new load will materialize and when, it is clear that substantial load growth is about to arrive15. The crux of the matter now is whether load-serving entities, regulatory bodies, and supply chains enable load growth at rates that jeopardize resource adequacy or introduce unreasonable power price volatility.

It is clear that many factors are at play when considering the impact of load growth on wholesale power prices. As you assess changing power market conditions and examine new power price forecasts, here are a few key points to focus on:

  • What assumptions are made about load growth and interconnection rates?
  • How quickly is new generation capacity built?
  • What kinds of generation capacity are built and where?
  • Which existing generation resources are kept online or reactivated?

While it is true that the power sector is living through a period of increased uncertainty and risk, not every worrisome headline should be taken at face value. By considering the above questions and market dynamics, buyers and sellers of clean power can ensure they’re making strategic decisions based on reality – not just hype.

This article contains a collection of information related to REsurety, Inc. and the commodity interest derivatives services and other services that REsurety, Inc. provides. Any statements of fact in this article are derived from sources believed to be reliable, but are not guaranteed as to accuracy, nor do they purport to be complete. No responsibility is assumed with respect to any such statement, nor with respect to any expression of opinion which may be contained herein. The risk of loss in trading commodity interest derivatives contracts can be substantial. Each investor must carefully consider whether this type of investment is appropriate for them or their company. Please be aware that past performance is not necessarily indicative of future results. Image: iStock / Petmal.


1 Multibillion-dollar data center projects to watch – Construction Dive; January 2025
2 U.S. Semiconductor Ecosystem Map – Semiconductor Industry Association
3 2024 United States Data Center Energy Usage Report – Lawrence Berkeley National Laboratory; December 2024
4 Electricity Demand Growth and Data Centers: A Guide for the Perplexed – Bipartisan Policy Center; February 2025
5 Powering Intelligence: Analyzing Artificial Intelligence and Data Center Energy Consumption – Electric Power Research Institute; May 2024
6 Queued Up: 2024 Edition Characteristics of Power Plants Seeking Transmission Interconnection As of the End of 2023 – Lawrence Berkeley National Laboratory; April 2024
7 Examples include: land acquisition, permitting, debt financing, availability of key components like transformers, skilled labor, and raw materials.
8 The Future of Resource Adequacy – U.S. Dept. of Energy; April 2024
9 PJM capacity prices hit record highs, sending build signal to generators – Utility Dive; July 2024
10 2024 Long-Term Reliability Assessment – North American Electric Reliability Corporation; December 2024
11 Georgia Power to Keep Coal, Gas Power Plants Running Longer as Demand Climbs – POWER Magazine; February 2025
12 The hottest trend in nuclear power: Reopening shuttered plants – Canary Media; September 2024
13 What DeepSeek Means for AI Energy Demand – Heatmap News; February 2025
14 DeepSeek might not be such good news for energy after all – MIT Technology Review; January 2025
15 Electricity Demand Growth and Data Centers: A Guide for the Perplexed – Bipartisan Policy Center; February 2025

Contact Us to Learn More

"*" indicates required fields

Name*
This field is for validation purposes and should be left unchanged.

Return to the blog post main menu.