Blackouts Could Increase 100x, DOE Reliability Report Finds

Retirements, Demand Growth, and Outages,

Retirement kills more people than hard work ever did- Malcolm Forbes

In April of 2024, the U.S. Department of Energy (DOE) under then-Secretary Jennifer Granholm released a report called The Future of Resource Adequacy, where the DOE argued that:

“Investing in all technology solutions, including clean energy generation and storage, transmission expansion and enhancement, and efficiency and demand management tools can provide ample, reliable and secure power in an age of rising electricity demand without having to rely on older, dirtier technologies.”

After four years of having a DOE that indulged in the childish fantasy that America can shut down its reliable coal and natural gas plants and rely on wind, solar, and battery storage to meet surging electricity demand, it’s clear the energy adults are now back in charge.

On Monday, the U.S. Department of Energy released a 73-page report entitled Evaluating the Reliability and Security of the United States Electric Grid, which concludes the United States will experience a 100-fold increase in blackouts if coal and natural gas plants are retired amid rising demand from data centers.

Method to the Madness

The report examined future grid reliability using multiple scenarios for power plant retirements in several regions throughout the country.

These regions, which are from the North American Electric Reliability Corporation (NERC) Integrated Transmission and Capacity Systems (ITCS) region, tend to split up RTOs into several regions, which you can see in the graph below.

Retirement Party

One retirement scenario, the “Plant Closures” scenario, allows the retirement of 104,000 MW of power plant capacity that have announced the intention to retire, and adds 209,000 MW of Tier 1 resources to the grid to replace them using data compiled by NERC.¹

A second scenario, the “No Plant Closures,” assumes the existing plants that have announced their retirements will not do so, and instead will remain online. Additionally, 209,000 MW of Tier 1 resources will come online to supplement the grid's supply.

In each of these scenarios, the Tier 1 resources are overwhelmingly wind, solar, or battery storage. The graph below shows that only 20,000 MW of new natural gas capacity is expected to come online by 2030, along with 31,000 MW of additional 4-hour batteries, 124,000 MW of new solar, and 32,000 MW of incremental wind.

Figure edited for clearer numbers and scenario names.

The graph also shows that the amount of coal capacity in the Plant Closures scenario would drop by nearly 72,000 MW, and natural gas capacity would fall by almost 5,000 MW.

At the same time, data center demand is expected to explode in the next five years, and these growth projections are incorporated into DOE’s electricity demand modeling to assess the impact of the rising power demand on grid reliability.

Data Center Demand Growth Is Big

DOE expects electricity demand for data centers to increase by 52,000 MW by 2030, representing about 6.7 percent of the current average peak demand in the United States. DOE’s estimate is a midpoint of several studies on the subject, ranging from 33,000 MW of demand growth estimates from Lawrence Berkeley Labs to 109,000 MW of expected demand growth from S&P.

The regional concentration of the data center demand growth is also worth noting. To allocate the 52 GW midpoint regionally, DOE “used state-level growth ratios from S&P’s forecast. These ratios reflect factors such as infrastructure, siting trends, and projected market activity.”

PJM appears poised to experience the largest growth in data center demand, followed by the West Non-California ISO, an area that includes virtually all of the Western Interconnect, excluding CAISO, and ERCOT. The Southeast (SERC), MISO, and SPP will see a more modest share of the data center demand growth.

However, the most interesting thing about the graph is that S&P is projecting zero demand increases in New England or New York, resulting from data centers. We don’t know why S&P made this projection, but we can speculate that it could be due to tight electricity supplies and consistently high prices in these regions.

Apart from data center demand growth, the DOE expects roughly an additional 50,000 MW of demand growth from other sectors of the economy, resulting in the expected peak demand for the country increasing from 774,000 MW to 889,000 MW in 2030, a 15 percent increase, or 2.3 percent annual growth. If not for data center demand the expected rate of growth would only be 1.1 percent.

Big, Bountiful, Blackouts

DOE modeled the impact of thermal retirements and rising demand in 2030 by using historical hourly fluctuations in wind, solar, and hydro generation and demand. Dispatchable plant capacity was calculated by adjusting the capacity on the grid in the NERC 2024 Long Term Reliability Assessment for historical outages.

For what it’s worth, this is how we would have approached the modeling, too.

DOE found huge blackouts throughout much of the country in the Plant Closures scenario, and far fewer blackouts in the No Plant Closures scenario.

The graph below shows SPP, ERCOT, parts of PJM South, and SERC East are at the highest risk of blackouts in both scenarios, which was likely influenced by the outage rates for natural gas during Winter Storm Uri (ERCOT and SPP) and Elliot (PJM and SERC East).

In the Plant Closures scenario, PJM would experience an average of 430 hours per year, resulting in 1.4 million MWh of unserved load, meaning blackouts would account for 0.15 percent of total generation, a metric called Normalized Unserved Energy (NUSE) in the report. SPP would also see an average of 379.6 hours of blackouts, with 313,797 MWh of unserved load, meaning 0.09 percent of demand would go unmet. Remember, grid providers are supposed to provide “five nines” reliability (99.999), and in this scenario, SPP would be only 99.91 percent reliable.

In a somewhat unexpected finding, the DOE report did not find capacity shortfalls in ISO New England or New York in either of the scenarios studied. This is likely due to the fact that load growth is expected to be small in both of these areas because there is no expected data center demand growth in these regions in the DOE study.

Conclusion: If You’re In a Hole Stop Digging

The DOE report presents a sober analysis of the current state of grid reliability, considering both announced retirements and the significant expected increase in electricity demand for data centers.

The report makes it exceedingly clear that most of the reliability problems we face in the coming years are the result of the unwise and unnecessary subtraction of reliable coal plants from the grid while replacing them with intermittent wind and solar generators that do not provide the same reliability attributes.

Combine the retirements, intermittent additions, and data center demand growth, and you have a recipe for disaster.

Beyond Based

Congrats to our friend

Emmet Penney

, who got a shoutout from The Man, Chris Wright, himself.

1. Tier 1 additions are those with signed interconnection agreements or power purchase agreements, or included in an integrated resource plan, indicating a high degree of certainty in their addition to the grid. Full details of the retirement and addition assumptions can be found in the methodology section of this report.

Commission won’t reconsider Rocky Mountain Power’s rate hike, calls request ‘offensive’ Utah News Dispatch.

Arguing that Utah ratepayers shouldn’t bear other states’ costs, the commission ordered a 4.7% electricity residential rate increase in April, only about a quarter of the 18.1% Rocky Mountain Power requested last August, and even less than the initial 30% proposal that upset Utah elected officials.

Meme credit Paul Phillips