Electricity reliability can mean how often customers lose power, how long interruptions last, how quickly the system is restored, or how outages affect businesses. The latest national distribution figures show 611.3 interruption minutes per U.S. customer in 2024 when major events were included, compared with 126.0 minutes when major events were excluded. These are different measures of the same year, not interchangeable estimates.
Contents
- National reliability metrics
- Outage scale and weather exposure
- Northeast and Illinois state comparisons
- Midwest and adjacent state comparisons
- Bulk-power-system events and restoration
- Business exposure and perceived importance
National reliability metrics
The U.S. Energy Information Administration’s Electric Power Annual Table 11.3 reports three distinct distribution-reliability indices for 2024. SAIDI measures interruption minutes per customer-year, SAIFI measures interruption events per customer-year, and CAIDI measures minutes per interruption. Keeping these definitions separate matters because a system can have frequent short interruptions or fewer but longer interruptions.
When all 2024 events were included, U.S. distribution customers experienced 611.3 interruption minutes per customer-year and 1.521 interruption events per customer-year. The average restoration duration was 401.9 minutes per interruption. EIA combines utilities that report under IEEE and non-IEEE methods, so the national values summarize the reporting population rather than a single uniform measurement system.
The major-event view changes the baseline substantially. Excluding major events, U.S. customers averaged 126.0 interruption minutes in 2024, 1.043 interruption events per customer-year, and 120.8 minutes per interruption. Major-event exclusions follow each utility’s reporting method, so the excluded-event comparison should be read as a reporting convention, not as a universal definition applied identically everywhere.
| 2024 U.S. distribution measure | Including major events | Excluding major events |
|---|---|---|
| Interruption duration | 611.3 minutes/customer-year | 126.0 minutes/customer-year |
| Interruption frequency | 1.521 events/customer-year | 1.043 events/customer-year |
| Restoration duration | 401.9 minutes/interruption | 120.8 minutes/interruption |
EIA’s 2024 reliability metrics represented an estimated 95.0% of customers, compared with 94.8% in 2023. Coverage is estimated from reliability-schedule meters divided by customers on the sales schedule. EIA notes that the construction can produce coverage above 100% for some states, another reason to treat state and national coverage as a reporting characteristic.
For a separate historical reference point, EIA reported that U.S. customers averaged approximately 5.5 hours of electricity interruptions in 2022. That EIA summary describes the multi-year pattern since 2013 as average interruption duration remaining around two hours per year after major events were excluded. The 2022 figure is an observed historical summary, not a 2026 measurement or forecast.
Outage scale and weather exposure
Weather can create a sharp difference between ordinary reliability and annual customer experience. The EIA reported 18 U.S. weather-related disasters that exceeded $1 billion in damage per event in 2022, citing NOAA/NCEI. Hurricane Ian left more than 2.6 million Florida customers without power in September 2022. Winter Storm Elliott left more than 1.5 million customers without power across several U.S. states in December 2022. Each is an event total, not an annual national average.
NERC’s 2025 State of Reliability overview uses a bulk-power-system footprint and a different event frame. NERC recorded 27 U.S. billion-dollar weather and climate events within that footprint in 2024, plus three Canadian events. It reported that 10 North American billion-dollar events had notable impacts on the bulk electric system. These counts should not be substituted for EIA’s distribution-customer minutes: they describe qualifying events and bulk-system impact rather than a customer-average interruption index.
The two perspectives answer different questions. EIA’s customer measures describe reported interruption duration and frequency across distribution customers. NERC’s event counts identify large weather and climate events affecting the bulk-power-system footprint. A year can therefore contain a relatively stable non-major customer baseline while still including high-consequence events that affect millions of customers.
Northeast and Illinois state comparisons
EIA’s 2024 any-method table provides a useful state-level comparison when major events were excluded. The values below are interruption minutes per customer-year, not total minutes for every customer and not a forecast.
| State | Minutes per customer-year |
|---|---|
| Rhode Island | 60.8 |
| Illinois | 59.4 |
| Massachusetts | 76.1 |
| New York | 72 |
| Connecticut | 73.1 |
| New Jersey | 99.7 |
| New Hampshire | 132.3 |
| Pennsylvania | 131 |
| Maine | 274 |
| Vermont | 275 |
Within this supplied set, Vermont and Maine had the highest reported averages at 275 and 274 minutes, respectively, while Illinois and Rhode Island had the lowest at 59.4 and 60.8 minutes. That comparison describes the reported 2024 state values under EIA’s table methodology; it does not establish a causal ranking or explain why states differ.
The table uses any-method reporting, and the state values exclude major events under the table methodology. EIA’s reporting coverage varies by state and can exceed 100% because of the numerator-and-denominator construction. The values are therefore best used as clearly labeled comparisons rather than as perfectly standardized measurements across every utility.
Midwest and adjacent state comparisons
The supplied 2024 observations also cover a group of Midwestern and adjacent states. All figures below use EIA’s interruption minutes per customer-year measure with major events excluded.
| State | Minutes per customer-year |
|---|---|
| Indiana | 121.4 |
| Michigan | 160.1 |
| Ohio | 133.2 |
| Wisconsin | 91.3 |
| Iowa | 89.8 |
| Kansas | 103.6 |
| Minnesota | 92.2 |
| Missouri | 106 |
| Nebraska | 73.6 |
| North Dakota | 84.2 |
In this group, Michigan reported 160.1 minutes per customer-year and Ohio 133.2, while Nebraska reported 73.6 and Iowa 89.8. Wisconsin, Minnesota, North Dakota, Kansas and Missouri fell between those values at 91.3, 92.2, 84.2, 103.6 and 106 minutes, respectively. These are state observations for 2024, not regional averages and not evidence that every utility within a state experienced the same reliability.
Comparisons are also limited by the EIA methodology. Utilities can report under IEEE or non-IEEE methods, and the major-event treatment follows the reporting utility’s method. The most defensible interpretation is therefore narrow: these are reported state-level interruption-minute figures for the stated year and exclusion rule.
Bulk-power-system events and restoration
Hurricane Helene illustrates why customer counts, infrastructure outages and restoration time should be reported separately. NERC recorded more than 4.7 million customers without power on September 27, 2024. It also reported approximately 431 transmission-element outages, the highest recorded for a single event in its cited comparison.
Functional transmission restoration was achieved 7.6 days after the first outage. NERC compared that duration with a 15-day average for previous Category 4 hurricanes and a nine-day average for all hurricanes in the preceding five years. Both comparisons are NERC’s cited prior-event averages for the bulk-power-system footprint.
Functional transmission restoration does not mean every end-use customer had power restored. That limitation is important when interpreting the 7.6-day figure: it measures restoration of transmission function, while customer restoration can depend on distribution damage and local conditions. The Helene figures are event-level observations from 2024, not an annual reliability average.
Business exposure and perceived importance
Electricity reliability also appears in firm-level survey responses. The World Bank study A Spatial Equilibrium Model with Endogenous Complementary Infrastructure Networks summarizes Enterprise Survey results in Table 2. The table’s single clean calendar survey year is not exposed in the supplied source excerpt, so the period is labeled as reported in the source; the table was accessed in 2026.
Among 1,802 surveyed firms, 45.6% reported experiencing power outages. The share was 45.2% among 814 small firms, 48.8% among 738 medium firms and 40.7% among 250 large firms. These are weighted firm-level survey observations, not household results and not utility-system averages.
Electricity was rated a very severe obstacle to operations by 24.4% of firms overall. The corresponding shares were 23.2% for small firms, 28.6% for medium firms and 36.5% for large firms. The larger-firm result is a reported perception in this sample, not a direct estimate of lost output or outage cost.
| Survey measure | Overall | Small | Medium | Large |
|---|---|---|---|---|
| Firms experiencing power outages | 45.6% | 45.2% | 48.8% | 40.7% |
| Electricity a very severe obstacle | 24.4% | 23.2% | 28.6% | 36.5% |
| Reliability extremely important when relocating | 24.9% | 24.2% | 27.9% | 30.1% |
Energy reliability was described as extremely important when relocating by 24.9% of firms overall, including 24.2% of small firms, 27.9% of medium firms and 30.1% of large firms. This is a firm-reported importance rating, not a revealed-location decision.
The same survey reported an average of 1.75 power outages in a typical month, with a reported standard deviation of 4.76. It reported an average power-outage duration of 1.74 hours, with a standard deviation of 5.15. These means describe sampled firms and are not directly comparable to EIA’s SAIFI or SAIDI measures; they use different populations, questions and measurement conventions.