Power Outages by State: How Long, How Often and Which States Are Getting Worse, 2015 to 2025

Power outages by state: West Virginia customers averaged 16.9 hours without power in 2025

In 2025 the average US electricity customer was without power for 262.6 minutes, or 4.4 hours. The five longest (West Virginia, Michigan, Nebraska, Alaska, Mississippi) were all at or above 501.2 minutes; the five shortest, led by the District of Columbia, were all at or below 117.5 minutes.

West Virginia customers averaged 1,013.2 minutes without power in 2025 (16.9 hours), the longest of the 50 states and DC, according to EIA Electric Power Annual 2025, Table 11.3. Michigan was second at 671.5 minutes (11.2 hours), followed by Nebraska (547.3), Alaska (540.9) and Mississippi (501.2).

The District of Columbia had the shortest outage time in 2025 at 50.3 minutes, followed by Arizona (69.9), Florida (70.5), Delaware (102.5) and Utah (117.5). The US figure of 262.6 minutes was 57.0% below the 611.3 minutes of 2024 and the lowest since 2016 (EIA Table 11.1).

SAIDI (System Average Interruption Duration Index): Total minutes of sustained interruptions per customer in a year, averaged over all customers a utility serves. EIA weights each utility's value by its customer count to build state and US figures. A customer is a metered account, not a person.

A single year can mislead, because one hurricane season can dominate a state's total. Over the 11 years from 2015 to 2025, Louisiana had the highest average at 1,110.3 minutes a year (18.5 hours), followed by Maine (1,073.1), West Virginia (854.4), South Carolina (658.2) and Mississippi (628.1), based on EIA Table 11.4. The District of Columbia had the lowest 11-year average, 68.7 minutes.

How long power outages last in the US: 196.3 minutes per interruption in 2025

Once the power went out in 2025, the average US customer waited 196.3 minutes (3.3 hours) for it to come back, or 114.7 minutes (1.9 hours) outside major events. Storm years stretch restoration time; on ordinary days it has stayed between 108.1 and 120.8 minutes since 2015.

How long power outages last depends mostly on whether a storm is involved. In 2025 the average interruption for a US electricity customer lasted 196.3 minutes (3.3 hours) including major events and 114.7 minutes (1.9 hours) excluding them (EIA Table 11.1). In 2024, the year of Hurricanes Beryl, Helene and Milton, the average interruption lasted 401.9 minutes (6.7 hours).

CAIDI (Customer Average Interruption Duration Index): SAIDI divided by SAIFI: the average number of minutes it takes to restore power once a sustained interruption starts. EIA reports it with and without major event days.

From 2015 to 2025, US restoration time including major events ranged from 164.5 minutes (2015) to 401.9 minutes (2024). Without major events it stayed between 108.1 and 120.8 minutes in every year, so an ordinary-day outage for the average customer lasts close to 2 hours (EIA Electric Power Annual 2025, Table 11.1).

By state, Michigan customers waited longest per interruption in 2025: 471.6 minutes (7.9 hours), followed by Nebraska (414.9), Missouri (335.1), West Virginia (328.8) and Pennsylvania (325.4). Arizona (78.3), Florida (82.2), Delaware (110.2), Nevada (113.6) and Louisiana (116.0) had the shortest (EIA Table 11.3). Michigan ranked second on total outage time while ranking 18th on frequency (1.424 interruptions), because each interruption lasted long.

Power outages in the US, counted per customer: the average customer had 1.338 sustained interruptions in 2025 and was without power for 262.6 minutes (4.4 hours) in total. Applied to the 153.9 million customers of utilities that reported reliability data, that is about 206 million customer interruptions in 2025, 162 million of them outside major events, against about 231 million in 2024 (ECW estimate from EIA data). A customer interruption counts each time one metered account loses power, so a single storm outage can add thousands.

Major events caused 67.5% of US outage minutes from 2015 to 2025

Outage minutes on ordinary days barely move from year to year; storm days drive the swings. In 2024, major event days produced 79.4% of US outage minutes; in the quieter 2025, 54.0%.

Without major event days, the average US customer lost 120.8 minutes of power in 2025 (2.0 hours) and 126.0 in 2024. Including them, the totals were 262.6 and 611.3 minutes (EIA Table 11.1). The difference is the time attributed to storms and other events outside normal operating conditions.

Major event day (MED): Under IEEE Standard 1366, a day on which a utility's daily SAIDI exceeds a threshold (Tmed) calculated from its previous five years of daily data. Utilities that do not use IEEE methods define major events themselves, according to EIA's table notes.

Across 2015 to 2025, major event days caused more than half of outage minutes in 36 of the 51 jurisdictions. The highest shares were in Florida (87.4%), Connecticut (84.0%), South Carolina (82.3%) and Louisiana (81.2%). The lowest were in the District of Columbia (21.4%), New Mexico (26.0%) and Wyoming (30.3%).

West Virginia is the clearest case of a state with long outages on ordinary days: 498.1 minutes without major events in 2025, the most of any state, and a major event share of only 50.8%. Mississippi (305.5 minutes) and Alaska (242.7) follow.

How often the power goes out: 1.338 interruptions per US customer in 2025

Frequency (SAIFI) and duration (SAIDI) rank states differently. Alaska had the most interruptions per customer in 2025 (3.203); the District of Columbia the fewest (0.324).

The average US customer had 1.338 sustained interruptions in 2025, down from 1.521 in 2024, and 1.053 excluding major event days (EIA Table 11.1). Alaska (3.203), West Virginia (3.082), Mississippi (2.594), Louisiana (2.324) and Maine (2.240) had the most (EIA Electric Power Annual 2025, Table 11.3).

SAIFI (System Average Interruption Frequency Index): The average number of sustained interruptions per customer per year, weighted by customers in the same way as SAIDI.

Outage time on ordinary days rose in 17 states and fell in 16 since 2015

Comparing 2015 to 2017 with 2023 to 2025, US outage minutes including major events rose 31.1%, but minutes without major events rose only 6.0%. Nationally, almost all of the growth came from major event days.

Using three-year averages to damp single-storm years, US SAIDI including major events went from 309.1 minutes (2015 to 2017) to 405.3 minutes (2023 to 2025), and rose in 37 of 51 jurisdictions (EIA Table 11.4). Without major events, the US average moved from 114.8 to 121.7 minutes.

On the ordinary-day measure, 17 states got worse by more than 10%, 16 improved by more than 10%, and 18 changed less than that. The largest rises were in Hawaii (+68.8%), Alaska (+62.3%) and California (+52.7%). The largest falls were in the District of Columbia (-67.2%), Wyoming (-38.9%) and Idaho (-35.0%).

Single-year comparisons tell the same story. US SAIDI without major events was 113.1 minutes in 2015 and 120.8 in 2025, a rise of 6.8%. US SAIFI without major events moved 0.7% over the same years, from 1.046 to 1.053.

DOE-417 disturbance reports fell to 426 in 2025 from 622 in 2024

Large incidents reported to the Department of Energy track the storm years. Reports for the 50 states and DC fell 40.5% in 2025, and incidents with loss of service to more than 50,000 customers fell 43.9%.

Utilities and grid operators must file Form DOE-417 for events such as a loss of electric service to more than 50,000 customers for 1 hour or more, physical attacks, cyber events and large load shedding (DOE-417 instructions). The DOE-417 annual summaries list 622 reports for 2024, the most of any year since at least 2013, and 426 for 2025.

Excluding events listed only for Puerto Rico or other territories, the 50 states and DC had 337 reports in 2025. Physical security events (vandalism, theft, suspicious activity including drones, and attacks) were the largest group at 131 (38.9%), followed by weather or natural disaster at 72 (21.4%), cyber at 42 and grid operations and supply at 17; 75 were other or unknown.

Reports of service loss to more than 50,000 customers, the criterion closest to what homeowners experience, numbered 64 in 2025 against 114 in 2024. The peak since 2015 was 125 in 2020. A DOE-417 report counts an incident, not customers or minutes, and a single storm can generate several reports from different utilities.

What the outage data means for standby generators and transfer switches

Standby generator demand follows outage years, and the installation work runs through electrical contractors. The public record supports a penetration figure but not state-level install counts.

Generac, which describes itself as the home standby product category leader, estimated in its fiscal 2025 10-K that home standby units reach approximately 6.75% of its addressable market, up from approximately 6.5% in its fiscal 2024 10-K. It defines that market as single-family detached, owner-occupied US homes valued over $175,000 (Generac 10-K, filed February 18, 2026). This is a company estimate, not an independent survey.

The same filing says the second half of 2025 had a very low level of baseline power outage activity, and it attributed lower domestic home standby and portable generator sales in 2025 to a significantly lower power outage environment and a strong 2024 comparison that included multiple major landed hurricanes. That matches the EIA data: US outage minutes dropped from 611.3 in 2024 to 262.6 in 2025.

Generac's 10-K describes its residential standby generators (7.5 to 150 kW) as permanently installed with an automatic transfer switch, and says its residential dealer network is made up of electrical and HVAC contractors who sell, install and service them. Connecting a generator or battery system to a home's wiring is licensed electrical work; homeowners can find a contractor through hire an electrician, and battery backup is covered in EV, solar and low voltage.

For contractors, the state ranking is a rough guide to where outage exposure is highest. West Virginia, Mississippi and Alaska combine long ordinary-day outages with high frequency; Florida, Louisiana and the Carolinas see most of their outage time in storm years. Contractor counts by state are in electrical contractor counts by state.

How we ranked the states

Get the calculations (CSV)

State by state

1. US customers averaged 262.6 minutes without power in 2025 including major events, and 120.8 minutes excluding them (EIA, any method).^[3]
2. US outage minutes fell 57.0% from 2024 (611.3 minutes) to 2025, the lowest level since 2016.^[3]
3. West Virginia customers averaged 1,013.2 minutes (16.9 hours) without power in 2025, the most of the 50 states and DC.^[1]
4. West Virginia also had the longest outage time without major events in 2025, 498.1 minutes, against 120.8 for the US.^[1][3]
5. The District of Columbia had the shortest outage time in 2025, 50.3 minutes per customer, followed by Arizona at 69.9.^[1]
6. Louisiana had the highest average outage time over 2015 to 2025, 1,110.3 minutes (18.5 hours) a year.^[2]
7. Major event days produced 79.4% of US outage minutes in 2024 and 54.0% in 2025.^[3]
8. Florida had the highest major event share over 2015 to 2025, 87.4%, but only 10.1% in 2025, when its total fell to 70.5 minutes.^[1][2]
9. Between 2015 to 2017 and 2023 to 2025, US outage minutes including major events rose 31.1% and minutes without major events rose 6.0%.^[2]
10. Ordinary-day outage minutes rose more than 10% in 17 states and fell more than 10% in 16 between 2015 to 2017 and 2023 to 2025.^[2]
11. Hawaii's outage minutes without major events rose 68.8% between 2015 to 2017 and 2023 to 2025, the largest rise of any state.^[2]
12. Alaska customers had 3.203 interruptions in 2025, the most of any state; the US figure was 1.338.^[1][3]
13. DOE-417 annual summaries list 426 disturbance reports for 2025 and 622 for 2024.^[10]
14. Reports of service loss to more than 50,000 customers in the 50 states and DC fell from 114 in 2024 to 64 in 2025.^[10][11]

In short

• The average US electricity customer was without power for 262.6 minutes (4.4 hours) in 2025, 57.0% less than the 611.3 minutes of hurricane-heavy 2024 (EIA Electric Power Annual 2025).
• West Virginia had the longest outage time in 2025 at 1,013.2 minutes per customer, followed by Michigan (671.5) and Nebraska (547.3); the District of Columbia had the shortest at 50.3.
• Over 2015 to 2025, Louisiana (1,110.3 minutes a year) and Maine (1,073.1) had the highest average outage time.
• Major event days accounted for 67.5% of US outage minutes from 2015 to 2025 and for more than half in 36 of 51 jurisdictions.
• Without major events, US outage minutes rose only 6.0% between 2015 to 2017 and 2023 to 2025, while minutes including major events rose 31.1%.
• On ordinary days, 17 states got worse by more than 10% and 16 improved by more than 10%; Hawaii, Alaska and California had the largest rises.
• US customers averaged 1.338 interruptions in 2025; Alaska had the most (3.203).
• DOE-417 disturbance reports fell to 426 in 2025 from 622 in 2024.

Common questions

Keep reading

• How to hire an electrician
• EV, Solar and Low-Voltage guide
• Solar Battery Cost 2026: Installed Price per kWh From Program Records
• Solar installer salary by state: BLS pay, jobs and outlook
• Solar panel recycling in the US: state laws, EPA's pending rule and how much panel waste is coming
• All electrical contractor data

How we built this report

Source selection: state and national reliability figures come from the U.S. Energy Information Administration (EIA) Electric Power Annual 2025, Tables 11.1, 11.3 and 11.4, released October 21, 2026. They are built from Form EIA-861 utility reports; the final 2025 EIA-861 files were released October 7, 2026. The 'any method' series is used throughout because it combines utilities that follow IEEE Standard 1366 with those that do not, covering 94.7% of US customers in 2025. Incident counts come from the U.S. Department of Energy DOE-417 annual summaries for 2013 to 2025, downloaded from doe417.pnnl.gov. Generac's 10-K filings are used only for a dated, company-stated penetration figure.

Inclusion and exclusion: the ranking covers the 50 states and the District of Columbia. Puerto Rico and other territories are excluded from rankings and from the DOE-417 '50 states and DC' counts, but Puerto Rico is described from EIA's own analysis. Momentary interruptions are not part of SAIDI or SAIFI. Market estimates and third-party rankings are cited only as pages compared against.

Validation: the 2025 state values were rebuilt from the utility-level file Reliability_2025.xlsx by weighting each utility's SAIDI by its reported customers, using the IEEE value where reported and the utility's own value otherwise. The rebuild came within a median 0.5% of EIA's Table 11.3 values, and within 2% for 38 of 52 jurisdictions; the rebuilt US figure was 260.9 minutes against EIA's 262.6. The 'State Totals' tab inside the same workbook covers fewer customers for some states (for example 369.1 minutes for Missouri across 1,472,695 customers, against 457.2 minutes in Table 11.3), so the Electric Power Annual tables are used.

Conflicting numbers: EIA's December 2025 article gives 11 hours for 2024, based on IEEE reporters (662.6 minutes, 11.0 hours); the any-method figure used here is 611.3 minutes (10.2 hours). The article's 80% major event share for 2024 matches the IEEE series (80.1%); the any-method share is 79.4%. EIA's 'nearly 53 hours' for South Carolina in 2024 compares with 3,136.5 minutes (52.3 hours) in Table 11.3. Both are shown where relevant.

Derived metrics: hours = SAIDI / 60; major event share = (SAIDI with minus SAIDI without) / SAIDI with; trend = change between three-year averages for 2015 to 2017 and 2023 to 2025, so that one storm year does not decide a state's direction; a state is 'worse' or 'better' when its ordinary-day SAIDI moved more than 10%. Every formula and input row is in derived_metrics.csv, and scripts/ rebuilds all files from raw/. Restoration time per interruption is EIA's CAIDI (SAIDI / SAIFI); hours = minutes / 60. US customer interruptions = US SAIFI x customers of utilities reporting reliability data in the EIA-861 file (ECW estimate, added 2026-11-03).

DOE-417 handling: the original OE-417 host (oe.netl.doe.gov) did not resolve on October 27, 2026; an Internet Archive copy of its summaries page from December 9, 2024 lists years through 2023, and the same annual files, plus 2024 and 2025, are now served by the DOE-417 site. A report is counted when its row has a valid start date. Cause groups (weather, physical security, grid operations and supply, cyber, other) are an ECW keyword mapping of DOE's free-text event types. The 2024 and 2025 files use a new layout with more categories, including cyber and drone sightings, so cause counts before and after 2024 are not strictly comparable. EIA's Electric Power Monthly Table B.1, which republishes these events, lists only January 1 to March 30, 2024 and was not used for counts.

Limitations: SAIDI and SAIFI are averages across customers; many customers had no outage and some had far more. Coverage varies by state, from 74.1% of customers in Nebraska to near complete in others. Major event days are defined per utility, so a storm counts differently across utilities. A 'customer' is a metered account, not a person. Nothing here is engineering, financial or safety advice.

Last updated: October 27, 2026. Next update: after the Electric Power Annual 2026 (expected October 2027) or the 2026 DOE-417 annual summary, whichever comes first.

This report presents published data with its limitations. It is not legal, financial or investment advice.

Sources by tier

Tier 1: primary and official sources (14)

• [1] U.S. Energy Information Administration. Electric Power Annual 2025, Table 11.3: Reliability metrics using any method of U.S. distribution system by state, 2025 and 2024. 2024 and 2025 (released 2026-10-21). Accessed 2026-10-27.
• [2] U.S. Energy Information Administration. Electric Power Annual 2025, Table 11.4: SAIDI values of U.S. distribution system by state, 2015 to 2025. 2015 to 2025. Accessed 2026-10-27.
• [3] U.S. Energy Information Administration. Electric Power Annual 2025, Table 11.1: Reliability metrics of U.S. distribution system. 2015 to 2025. Accessed 2026-10-27.
• [4] U.S. Energy Information Administration. Electric Power Annual (release date October 21, 2026; next release October 2027). 2025 data year. Accessed 2026-10-27.
• [5] U.S. Energy Information Administration. Annual Electric Power Industry Report, Form EIA-861, 2025 final data (Reliability_2025.xlsx). 2025. Accessed 2026-10-27.
• [6] U.S. Energy Information Administration. Annual Electric Power Industry Report, Form EIA-861, archived data files 2013 to 2024. 2013 to 2024. Accessed 2026-10-27.
• [7] U.S. Energy Information Administration. Hurricanes in 2024 led to the most hours without power in the United States in 10 years (Today in Energy, December 1, 2025). 2024. Accessed 2026-10-27.
• [8] U.S. Energy Information Administration. Duration of power outages in Puerto Rico not caused by major events increased 19% in 2025 (Today in Energy, August 10, 2026). 2021 to 2025. Accessed 2026-10-27.
• [9] U.S. Energy Information Administration. Average frequency and duration of electric distribution outages vary by states (Today in Energy, April 5, 2018). 2016. Accessed 2026-10-27.
• [10] U.S. Department of Energy, Office of Cybersecurity, Energy Security, and Emergency Response. DOE-417 Electric Emergency Incident and Disturbance Report, annual summaries 2013 to 2025. 2013 to 2025 (2024 and 2025 files generated 10/7/2026). Accessed 2026-10-27.
• [11] U.S. Department of Energy. Form DOE-417 instructions (OMB No. 1901-0288, expires 05/31/2027). current form. Accessed 2026-10-27.
• [12] U.S. Department of Energy (archived by the Internet Archive). Electric Disturbance Events (OE-417) Annual Summaries page, Internet Archive copy of December 9, 2024 (original host unreachable on 2026-10-27). 2000 to 2023 listings. Accessed 2026-10-27.
• [13] U.S. Energy Information Administration. Electric Power Monthly (data for August 2026, released October 23, 2026), Table B.1: Major disturbances and unusual occurrences. January 1 to March 30, 2024. Accessed 2026-10-27.
• [16] U.S. Energy Information Administration. Annual Electric Power Industry Report (EIA-861 data file) page: 2025 final data released October 7, 2026; next release August 2027 (early release). 2001 to 2025. Accessed 2026-10-27.

Tier 2: trade bodies, market research, company data (2)

• [14] Generac Holdings Inc. via SEC EDGAR. Generac Holdings Inc. Form 10-K for fiscal year 2025 (filed February 18, 2026). fiscal year 2025. Accessed 2026-10-27.
• [15] Generac Holdings Inc. via SEC EDGAR. Generac Holdings Inc. Form 10-K for fiscal year 2024 (filed February 19, 2025). fiscal year 2024. Accessed 2026-10-27.

Tier 3: journalism and expert commentary (4)

• [17] MRO Electric. U.S. power outages by state (June 28, 2024). 2018 to 2022. Accessed 2026-10-27.
• [18] ROI-NJ. N.J. ranked 25th among states most at risk for power outages in 2025 (June 3, 2025). May to August, 2019 to 2023. Accessed 2026-10-27.
• [19] Government Technology. Which state experiences the most power outages?. 2018 to 2022. Accessed 2026-10-27.
• [20] EC&M. U.S. States with the Most Power Outages (March 9, 2021). EIA data as reported by MRO Electric. Accessed 2026-10-27.