Earlier this month, Advanced Energy United, alongside the American Clean Power Association and the American Council on Renewable Energy, requested rehearing of the U.S. Department of Energy’s (DOE) Resource Adequacy Report, published July 7, 2025, arguing that DOE “missed the opportunity to present all the viable types of energy needed to address reliability and keep energy affordable.”

We took this step because, if left unchallenged, DOE’s findings could shape grid policy in a way that locks consumers into paying more for outdated approaches and prevents existing power plants from retiring, while sidelining proven advanced energy-based solutions.

Background: What Prompted DOE’s Report? 

DOE’s release followed Executive Order (EO) 14262 (Strengthening the Reliability and Security of the United States Electric Grid), directing DOE to evaluate the “reliability and security” of the United States energy grid during an “unprecedented surge in electric demand” from artificial intelligence (AI) and data center development, and to develop a protocol to expedite orders under Section 202(c) of the Federal Power Act (FPA). Section 202(c) is meant to direct temporary connection, generation, delivery, or transmission of electricity to meet a sudden increase in demand, reduction in supply, or other identified emergency.

In contrast, DOE was directed to report on the potential future resource shortfalls, identify risks due to generator retirement and increased demand, and then use these findings to prevent existing resources in at-risk regions from leaving the grid. The EO did not direct DOE to consider broader solutions like new resource development, faster interconnection, virtual power plants, or load management.

DOE’s analysis was not intended, and should not be understood, as merely a “report,” but rather a protocol for pursuing a preferred solution to addressing resource adequacy needs. In this protocol, DOE predicts significant blackout risks in 2030, which it blames on retirement of existing generation outpacing firm, dispatchable additions; AI-driven demand growth; and over-reliance on renewable generation.

Report Shortcomings 

Our rehearing request details several critical flaws with these findings and with the process leading up to the protocol’s release.

First, although DOE labeled this as a “report,” it qualifies as a rule under the Administrative Procedure Act, making it subject to rehearing challenges. Procedural flaws include: issuing the protocol without a notice-and-comment period, failing to identify any qualifying “emergency” to justify action under 202(c), exceeding statutory authority by effectively claiming regulatory control over resource adequacy (which is reserved for the states and other federal regulators), and disregarding the Information Quality Act’s requirements for peer review, replicability, public input, and use of the best available data. 

Substantively, the protocol’s findings are also flawed as a result of DOE’s unrealistic and internally inconsistent assumptions. In short, the protocol fails to conduct scenario-based analysis, instead picking edge-case inputs that lead to an unrealistic, worst-case projection of future risk. Specifically, the protocol:

• Assumes implausibly low resource additions. The protocol assumes just 209 GW of new resource entry between now and 2030, despite nearly 2,000 GW of resources waiting in interconnection queues. This is because DOE relies on only very late-stage projects, and openly admits that it assumes virtually no resource entry after 2026—a particularly unrealistic assumption in light of the high load growth numbers the report also assumes. There are no sensitivities or alternative cases that consider higher resource entry.

• Includes high resource retirements. The protocol includes a single retirement scenario, which assumes roughly twice as many resource retirements by 2030 relative to recent evaluations by the North American Electric Reliability Corporation (NERC) and the U.S. Energy Information Administration. 

• Misrepresents load growth. The protocol is in good company in assuming high load growth from data centers and AI, among other drivers, but includes only one load growth scenario that does not capture uncertainty, and ignores opportunities to manage this growth. The protocol also acknowledges that resource planners would not allow load growth to jeopardize reliability, directly contradicting the report’s topline conclusion. Demand growth that outstrips supply is more properly described as an economic development opportunity cost than a reliability risk.

• Overlooks state and RTO planning tools. States and RTOs—the entities with responsibility to ensure resource adequacy—rely on planning tools and market signals to course-correct when shortfalls arise. The report ignores these measures altogether.

• Understates winter reliability risks for thermal generation. DOE relies on historical datasets that include extreme weather events that exposed the vulnerabilities of the thermal resource fleet—yet reaches the unsupported conclusion that more of these resources would prevent blackouts.

• Disregards the resource adequacy benefits of interregional transmission. Connections between grid regions can deliver significant resource adequacy benefits that the report does not consider at all, despite relying on the data and methodology of a recent NERC report that identified 35 GW of “prudent” interregional transfer capability needs.

What DOE’s Approach Really Means 

It should be no surprise that a report assuming very low resource entry, very high resource exit, and significant load growth, will spit out a “red alert” result. But the report is not intended as a thought experiment to identify the very worst-case scenario; it’s a tool explicitly tasked with preventing power plant retirements. We have already seen DOE take action under Section 202(c) to forestall upcoming plant retirements (e.g., Campbell in MI and Eddystone in PA). This approach risks propping up plants where closure may be justified, and risks undermining more affordable, durable solutions to the larger, long-term demand crisis.

As a case in point, a recent study by Grid Strategies titled “The Cost of Federal Mandates to Retain Fossil-Burning Power Plants” found that keeping old, inefficient fossil fuel plants running beyond their useful life would cost consumers more than $3 billion per year. This analysis looked at plants that are scheduled to retire by 2028 and others that may announce retirements in an effort to receive ratepayer subsidies to stay open, given the DOE’s willingness to hand out subsidies to fossil plants, whether they are needed for reliability or not. These plants are set to retire because they are not needed and not able to compete with newer, more efficient generation. In most cases, their retirements were approved by either state regulators, RTOs/ISOs, or both. The study by Grid Strategies estimates that if additional fossil fuel power plants move up their retirement dates in response to this DOE protocol, ratepayer costs could reach nearly $6 billion annually.

Solutions for A Reliable, Affordable Grid 

Ultimately, DOE’s resource adequacy report fails to provide the kind of measured assessment of the grid’s future that any resource planning decisions should be based on and ignores available solutions to meet the challenges posed by rapid load growth. Properly crediting renewable resources for their contribution to grid reliability, and faster interconnection of all resources to the grid, are among the solutions DOE could easily encourage, instead of subsidizing old, inefficient plants that cost ratepayers and provide few benefits.

Given the express intent to rely on this report to prevent power plant retirements, DOE’s protocol could play a major role in resource outcomes—all without consideration of the cost implications, appropriate consultation with states and grid operators, and with no evaluation of alternatives.