Brief

AT A GLANCE
As the nuclear energy sector prepares for its first major expansion in a generation, targeted workforce development provides the foundational talent needed to successfully drive this historic growth. The federal government has set a goal of quadrupling U.S. nuclear capacity to 400 gigawatts by 2050.1 Vogtle Units 3 and 4 entered service in 2023 and 2024, three closed plants in Michigan, Pennsylvania, and Iowa are being restarted, and advanced reactors are in development. Construction of Vogtle Units 3 and 4 required more than 9,000 craft workers at its peak, the four government-owned plants with disclosed county employment have approximately 450 to 700 employees per reactor, and reactor operators complete approximately two years of training before NRC licensing. DOE's Pathways to Commercial Liftoff report estimates that deploying 200 gigawatts of new nuclear capacity by 2050 would require about 375,000 additional workers, about 100,000 to operate the new reactors and about 275,000 to construct them and manufacture components, compared with about 100,000 workers today.2
Planning for that growth starts with an accurate picture of the workforce already in place: how many people work in nuclear energy, in which occupations and states, what they are paid, and how many are approaching retirement. This brief assembles that picture from publicly available data, with 2024 as the baseline year and 2025 values where sources report them, drawing on federal statistical sources, including DOE's U.S. Energy and Employment Report, company filings, and NRC records. The technical supplement contains the full analysis, and the appendix sets out the methods and the status of each source.3
This is the first brief in a series. It uses publicly available data to provide a baseline for further research. Later briefs cover the policy landscape, education and training pathways, workforce demand, and the labor needs of different parts of the industry.
The BLS census count of plant employment and the DOE survey count overlap, and neither contains the other. In this brief, plant employment means employment at establishments classified as nuclear electric power generation (NAICS 221113), mostly plants plus some corporate and administrative offices. The survey's 67,900 comprises 57,900 in electric power generation, of which 40,100 are in utilities, and 10,000 in the fuel cycle. The census count of 42,260 at nuclear generating establishments is 2,160 higher than the survey's utilities figure, so the counts cannot be added or subtracted. The radiation monitoring count includes people on no nuclear employer's payroll, mainly contractors mobilized for outages, and it also covers plant employees and workers at licensees other than commercial reactors, such as industrial radiographers, fuel processing facilities, and byproduct material users. It is the only public count that includes the contract outage workforce, which no employment series is designed to count, and it does not isolate that workforce.5 DOE's estimate of the workers expansion will need uses a baseline of about 100,000 today, which matches none of the three counts, so it cannot be compared directly with any of them.
Several groups fall outside or only partly inside these counts. New-build construction is project-based and may not appear in year-to-year counts: Vogtle Units 3 and 4 employed more than 9,000 craft workers at peak and demobilized after Unit 4 entered service in 2024. The Idaho National Laboratory employed about 6,000 people across all its programs in fiscal 2023, with roughly half its budget from DOE's Office of Nuclear Energy. GAO reports that the NRC lost about 500 staff on net between July 2024 and June 2026.6 Advanced reactor developers are classified with research and engineering firms and are not separately measured. Headcounts are available for five developers and total about 3,700: two from company filings (NuScale, 428, and Oklo, 205) and three from third-party estimates (TerraPower, X-energy, and Kairos Power).7 Naval propulsion, the weapons complex, DOE environmental cleanup, and medical isotope production share training pipelines with civilian nuclear energy but are excluded.
Employment at nuclear generating establishments fell from 51,592 in 2014 to 40,221 in 2021, a decline of 22 percent, over a period in which seven plants closed. From 2016 to 2021 the census count fell 19.7 percent and the survey count of workers attributed to nuclear generation fell 18.5 percent, so the two sources agree on the size of the decline. Plant employment has been flat since 2021, apart from an increase in 2024 that does not reflect hiring: federal employment in the industry rose by 1,964, close to the combined employment of TVA's Watts Bar and Sequoyah plants, which appear in the industry data for the first time that year.8
In 2025, federal data shows little change at plants. Employment at nuclear generating establishments rose 0.5 percent in 2025, with private employers adding about 180 workers. Company filings point to faster growth outside plants: fuel supplier Centrus grew 45 percent and reactor developer Oklo 81 percent in 2025. DOE's 2026 survey, published in September 2026, reports that nuclear employment rose 4 percent, or 2,300 workers, in 2025, which, set against the flat plant count, suggests growth outside plant operations.9
The occupational mix changed more than the total. Reactor operators and nuclear technicians fell by 17 and 19 percent, in line with the 20 percent decline in plant employment, which is consistent with NRC staffing requirements that tie licensed operator numbers to operating reactors. Nuclear engineers fell by 49 percent. Nationally, employment of nuclear engineers across all employers fell 31 percent between 2016 and 2022 and then recovered 25 percent by 2025, with the growth at engineering services and research and development firms rather than at plants. About one nuclear engineer in four worked at a nuclear plant in 2023 to 2025, down from about two in five in 2015 to 2017.11 Reactor developers are classified in the industries where that growth occurred, so some of it is plausibly developer hiring, though the data cannot identify employers. The share of plant employment held by security officers rose from 9.2 to 10.5 percent, and they are now the second-largest occupation at plants, close behind reactor operators.
Occupational data shows the decline at plants and the shift in engineering. Other sector dynamics do not discernibly alter these estimates but help contextualize them. Some hiring does not measurably shift estimates at the sector level. Vogtle Units 3 and 4 added roughly 800 permanent positions, about 2 percent of plant employment and too few to distinguish from the error in the estimates.12 Retirements that are replaced also leave occupation counts unchanged. Uprates and license extensions sustain employment rather than add to it.
The state figures come from the DOE survey and are allocations rather than counts, and federal census data for most states is withheld to protect employer confidentiality. On the survey's measure, 95 percent of nuclear generation employment is in the 28 states with operating reactors, led by South Carolina, Illinois, Pennsylvania, California, and New York. Several new-build and advanced reactor projects are in states with little nuclear generation employment on this measure, including Wyoming and Idaho, although Idaho hosts the national laboratory, which the generation figures exclude.
The median wage at nuclear generating establishments was $118,000 in 2024. The median for the electric power generation workforce was $65,430 and the national median was $49,500, so the nuclear figure is 80 percent and 138 percent higher. The nuclear figure is a BLS estimate for plant establishments and the comparison figures are DOE survey estimates for the whole generation sector, so the comparison is indicative. Figure 3 illustrates the difference by occupation. DOE's summary of its 2024 survey reports a premium of about 50 percent, citing an NEI analysis.14 The premium persists within shared occupations: electricians at nuclear plants earned 48 percent more than electricians across the sector, and electrical engineers 28 percent more. For reactor operators, nuclear engineers, and nuclear technicians, the sector figure is the national median for the occupation, so the comparison carries little information.
The median wages above are straight-time wages. Average annual pay at private nuclear generating establishments in the census data, which includes overtime, bonuses, and other pay beyond base wages, was $166,931 in 2024, 37 percent above the BLS mean wage of $121,550 for the same establishments.16 The wage estimates exclude overtime and bonuses, and the data do not show how much each component contributes to the difference. Survey wage figures therefore understate total pay at these establishments.
Real wages diverged by type of work. Between 2015 and 2025, median wages at nuclear plants for electricians, industrial machinery mechanics, and production supervisors rose 12 to 14 percent after inflation, while those for reactor operators, nuclear technicians, and electrical engineers rose 1 to 3 percent and those for nuclear engineers fell 3.9 percent.17 Rising real pay for trades is consistent with the shortage of skilled craft workers that industry reports describe. A decade of falling real pay for nuclear engineers is harder to reconcile with reports that engineers are among the hardest roles to fill, although a median can fall when senior engineers are replaced by junior hires and the wage estimates exclude bonuses. The brief does not compare nuclear wage growth with wage growth in fossil, solar, and wind generation; that comparison requires census pay data for those industries.
Pay varies by location, but public data cannot isolate the same role across places. Where 2024 census data is disclosed, average pay at private nuclear employers was $179,399 in Pennsylvania and $148,491 in South Carolina, and at four government-owned plants it ranged 23 percent, from $147,916 at Cooper in Nebraska to $181,651 at Sequoyah in Tennessee.18 These averages cover all occupations at each site, so they mix regional pay with differences in staffing and with corporate and administrative units, and they are unstable: Pennsylvania's average rose from $163,897 in 2023 to $179,399 in 2024 and $210,888 in 2025, while South Carolina's rose 5 and 4 percent. State estimates for reactor operators, nearly all of whom work at nuclear plants, would isolate same-role variation but are not used in this brief because they have not been retrieved.
Nuclear generation has the highest union representation of any generation subsector, 19 percent, against 13 percent for electric power generation and 7 percent for the private workforce nationally; nuclear fuels is at 18 percent. The survey measures employers with at least 20 percent of their workforce unionized, not the share of individual workers covered, and it excludes public employers.19 Union coverage follows occupational lines at individual sites, as collective bargaining agreements show: Centrus, for example, reports separate agreements with the United Steelworkers and with the Security, Police, and Fire Professionals of America, a union of security officers.20 The only nuclear-related turnover figure is enterprise-wide, about 4 percent a year at PG&E, and NEI reports rising voluntary attrition among workers in their first ten years.
Nuclear generation has the highest share of mid-career workers of any generation subsector (59 percent aged 30 to 54) and the lowest share of workers under 30 (24 percent, against 29 percent for electric power generation as a whole). Most of the 5-point difference between nuclear and the sector comes from solar and wind, whose workforces are young and growing and supply most of the sector's workers under 30. Against coal, natural gas, water power, and combined heat and power, nuclear's under-30 share is 1 to 4 points lower. DOE publishes no margin of error for these subsectors, and the only subgroup margin it reports is about 3 points. The share has been stable at 23 to 24 percent across the three editions that report it.
The survey does not publish a median age. An estimate from its age bands puts the average nuclear generation worker at about 41, compared with about 40 for electric power generation and for the energy workforce and about 42 for the national workforce. Nuclear's average is higher than the sector's even though its share over 54 is similar, because it has a smaller share under 30 and a larger share aged 30 to 54.22
Published descriptions of the workforce's age differ. DOE's Office of Nuclear Energy describes it as aging,23 and NEI's 2023 workforce plan describes it as younger and less experienced because retirees have been replaced, particularly in operations and engineering. INL describes plant staffing as concentrated among workers near retirement and workers in their first ten years.24 The descriptions rest on different measures. DOE's rests on the survey's age bands, which show the share aged 55 and over rising from 12 percent in 2016 to 18 percent in 2024. NEI's and INL's describe years of service, which the survey does not report and which can differ from age because new hires include people in their thirties and forties. Tenure data that would test the descriptions is not published.
Both descriptions can be accurate. Eligibility estimates of 35 to 39 percent measure who could retire, not who is old, and circulated for years, which may have shaped the perception of an aging workforce. The share aged 55 and over has risen, but the estimated average age of about 41 is close to the national 42, and NEI's account of a younger, less experienced workforce is consistent with the nearly 15,000 hires it reported for 2009 to 2012. The workforce may be aging in some places and younger in others, by plant, role, or years of service, but public data reports none of these breakdowns.
By the survey's measure, women and non-White workers are each 32 percent of nuclear generation workers, above the electric power generation sector, while Hispanic workers are underrepresented against every comparator and veterans are the lowest share of any generation subsector. The only public breakdown by role, from 2005 federal equal employment records, showed women at 15 percent of nuclear plant employees and 3 percent of craft workers, and minority workers at 10 percent, falling from about 20 percent in service roles to under 6 percent among managers.25 The two sources measure different populations by different methods, and the difference cannot be read as change over time. Neither source reports nuclear demographics by state or region.
Retirement exposure determines how many workers the nuclear sector must recruit and how quickly. Many nuclear roles require long qualification and plant-specific experience, and experienced staff are not readily replaced by new hires. Expansion adds to the replacement need, because the sector must fill positions left by retirements as well as new positions. At the NRC, voluntary retirements accounted for most of a net loss of about 500 staff between July 2024 and June 2026.26
Reported estimates of retirement exposure in nuclear energy come from industry surveys and from DOE's summary of its energy jobs survey. NEI's 2007 workforce survey of member utilities reported that up to 35 percent of the workforce could retire within five years, and the figure was cited at a 2008 NRC Commission meeting as 35 percent of staff at the 104 operating reactors being retirement-eligible by 2012. Later NEI estimates were 38 percent in 2012, described as set to retire within a few years, and 39 percent in 2015, described as eligible to retire by 2018. The 2007 and 2015 figures measure retirement eligibility among utility employees; they do not measure age, or how many eligible workers retire. The 39 percent estimate was still cited, without an updated date, in the annual reports of GSE Systems, a nuclear training and simulation company, through fiscal year 2022.27 NEI reported in 2013 that the industry had hired nearly 15,000 people between 2009 and 2012.28
DOE's Office of Nuclear Energy, summarizing the 2024 edition of the energy jobs survey, described the nuclear workforce as aging and expected a significant share to retire over the next decade, citing the 60 percent of nuclear workers aged 30 to 54 and the 23 percent smaller share of workers under 30 than in the energy workforce overall.29
Eighteen percent of nuclear generation workers were over 54 in 2024, against 17 percent for electric power generation, 19 percent for the energy workforce, and 24 percent nationally, up from 12 percent in 2016. One recent eligibility figure from a nuclear-operating utility, in PG&E's 2024 annual report, is 18 percent of employees across the company, with retirement age defined as 55 and older.30 Figure 6 shows annual retirements estimated by GEC's cohort model from the published age data, with projections to 2035; the model is described below.
A rising share of older workers does not show how many are leaving, because the group is replenished each year by workers turning 55. GEC built a cohort model that tracks the workforce by single year of age and estimates departures from the published age data. Figure 6 shows its estimates. It estimates about 930 retirements a year between 2017 and 2021, when employment was falling, and about 680 a year between 2021 and 2024. In 2017 to 2021 retirements were about 1.5 percent of the workforce a year and under a fifth of all departures; employment fell because hiring fell short of total departures. Part of the apparent fall in the retirement rate after 2021 may reflect a change in how the survey measured age, which coincides with a change in its publisher.
Holding recent rates constant, the model projects retirements rising from about 860 in 2025 to about 1,060 a year by 2028 to 2030 and then easing. That is an increase of about a quarter at the peak and roughly 10,000 retirements over the decade, or 17 percent of the 57,900 workers attributed to nuclear generation in 2024. The projection is conditional on current hiring and retirement patterns and is not a forecast.
These results are consistent with DOE's description: the share aged 55 and over has risen, the estimated average age is above the sector's, and the model estimates that about 17 percent of the 57,900 workers retire over the next decade. They differ from the NEI estimates in what they measure, since the model estimates retirements from age data and eligibility is a different quantity. The 18 percent aged 55 and over and PG&E's 18 percent eligible are lower than NEI's 35 to 39 percent, although the workforces and years differ.
Retirement exposure by occupation and region is not estimated here.
The baseline year is 2024, the latest year that the DOE survey, BLS census and occupational data, and inflation data all cover. Where a source reports 2025, as the BLS series and DOE's 2026 survey do for employment, the brief gives the 2025 value. DOE survey figures for age, wages, and demographics are for 2024.
Occupation codes alone cannot identify nuclear work, but industry codes can for plants. Only nuclear power reactor operators have an occupation code used solely in nuclear; nuclear engineers and technicians mostly work elsewhere, and every other occupation is shared with non-nuclear work. Every worker counted within the industry code for nuclear electric power generation is employed at an establishment classified as nuclear generation, which includes some corporate and administrative units.
Public data measures plant operations best. Outside plant operations, nuclear work shares its industry classification with non-nuclear work: in construction, contract craft, component manufacturing, engineering services, reactor development, decommissioning, and much of the fuel cycle. Public data therefore describes these segments less completely than plant operations.
No public count consistently includes contractors, and public sources do not report measured employment by state (DOE's state figures are allocations) or age, tenure, and union coverage by occupation. Retirement exposure by occupation and region therefore cannot be estimated, and which roles skew oldest or youngest cannot be determined.