← All Content & Research
Workforce Intelligence Brief · Nuclear Sector

America is about to build reactors it can't staff.

The U.S. has committed to roughly quadrupling nuclear power by 2050. To run those plants, the workforce must grow from about 100,000 today to 375,000. At the same time, 40% of today's workers can retire within ten years. The people don't exist yet, and each one takes two to four years to train. This brief shows how wide the gap is and what it takes to close it in time.

SOC 19-4051
BASE YEAR 2024
HORIZON 2030 / 2050
PREPARED June 2026
The official forecast
−8%
The Bureau of Labor Statistics expects nuclear technician jobs to fall from 6,000 to 5,500 by 2034. It assumes the fleet keeps shrinking, as it did for twenty years.
What the build-out needs
+275%
The new federal plan needs the nuclear workforce to nearly quadruple. Operators and technicians are the largest group that has to grow.
Bad
The verdict
Fixable, but only if we start now.

Demand is rising fast, the supply of trained people is thin, and training takes years. That combination is a lasting shortage, not a rough patch. Employers already feel it: 63% of nuclear power manufacturers called hiring "very difficult" in 2024, the worst score of any electricity sector. The good news is that the problem is well understood and the fixes are known. The only real question is when we begin.

Good
Average
Bad ◄ we are here
Crisis
Stable today · the fleet is staffed Start now → manageable · wait 3 years → crisis
If we start now

Manageable

A program that opens in 2026 graduates its first qualified people around 2028 to 2030. That is exactly when the new plants come online and the retirements peak. Supply meets demand.

If we wait

Crisis

Training time can't be rushed. Wait three years and the first graduates show up after demand has already spiked. The result is delayed projects, higher costs, and a scramble for scarce contract labor.

01

Why the official number is misleading

The first thing to settle is the contradiction: the government projects a decline while the policy calls for a boom. Both numbers are real. They just measure different futures, and only one of them describes the country we now live in.

How the two numbers reconcile

The −8% forecast looks backward. It models the world as it was in 2024, a shrinking and aging fleet with nothing new being built, and assumes that trend holds. It describes the status quo.

But the status quo ended. The federal orders of May 2025, more than $20 billion in private nuclear deals from data-center firms, and the first permits for small reactors all came after that forecast was set. Those events point the other way. When a plan to quadruple capacity meets a workforce projected to shrink, the gap between them is the shortage. Plan against the demand, not the old forecast.

63%
Of nuclear manufacturers called hiring "very difficult" in 2024, the worst of any electricity sector
DOE USEER 2025
~40%
Of the workforce can retire within ten years, and the wave is starting now
NEI / DOE
454
Nuclear engineering graduates a year, down 25% from a decade ago
ORISE 2022
$104,240
Median technician pay, among the five highest-paid roles in power generation
BLS, May 2024
02

Who is in these jobs today

Before measuring the gap, look at the people already doing the work. There aren't many of them, they skew older than most fields, and there is no bench behind them. The workforce runs fine as it is. It cannot grow quickly.

67,900
Total nuclear energy workers in 2024, flat against 68,008 the year before
DOE USEER 2025
~12K
Operators and technicians specifically, the largest group that runs a plant
USC Marshall / NCEA
85%
Of nuclear jobs are in power generation, not fuel production
DOE USEER 2025
~6,000
Workers in the narrow "nuclear technician" job code; most plant techs are counted elsewhere
BLS OEWS
How big is the workforce, really

It depends on how you count. The "100,000 workers" figure uses the widest definition. The operating fleet itself runs on about 58,000 to 68,000 people. The operators and technicians inside that group number in the low tens of thousands, and the narrowest job code captures just 6,000. However you slice it, this is a small, specialized workforce. That is exactly why a retirement wave hits so hard: there is no deep bench to absorb the losses.

Age: the retirement exposure
Under 30
~20%
30–60
~60%
60+
~20%
Roughly 60% of workers are between 30 and 60. The risk isn't only the 20% over 60. It's that a third to 40% become eligible to retire within five to ten years.
Who's missing: the untapped pool
32% WOMEN
Women in nuclear: 32%
Men in nuclear: 68%
U.S. workforce overall: 47% women
Women make up 47% of the U.S. workforce but only 32% of nuclear. Closing that 15-point gap is the single largest source of new workers, before counting Navy veterans or transfers from other industries.
02

The problem hits from two sides at once

Side 1 · The top

Retirements drain hard-won knowledge

Senior workers are retiring, and they take with them the unwritten know-how of how each plant actually behaves. The people who should replace them, the 10-to-15-year veterans, were never hired, because no plants were built for two decades. Seniors leave, juniors arrive, and there is no one in between.

~⅓ to 40% can retire within 5–10 years
Side 2 · The bottom

Young hires quit too soon

The young workers the industry does hire are leaving faster than before, often pulled away by renewables and tech. You can pour graduates into the bucket, but it has a hole in the bottom. Keeping people is as much the problem as training them.

Rising turnover under 10 years of service
02

So how bad is it, really?

On paper
Stable today. That's the trap. The running fleet is fully staffed, so nothing looks broken. That false calm is why the institutions that train this workforce feel no urgency, while the demand curve they serve is already bending up.
The institutions
Built to match, not to scale. The colleges and programs that produce nuclear technicians and operators size each class to today's jobs on purpose. They are not retooling to graduate the 3–4× more people the build-out needs. The supply side cannot meet the demand coming at it.
All at once
Demand spikes into a flat pipeline. Quadrupling capacity, ~40% of the workforce retiring, and young hires walking out all land at once on a training system that isn't expanding. The gap doesn't add. It compounds.
The hard limit
The clock has already started. A licensed operator takes 18–24 months that cannot be shortened, on top of years of schooling. If the workforce-development system doesn't start scaling now, the people simply will not exist when the reactors are ready for them.
The takeaway

So how bad is it? Bad. The problem is the supply side, not the demand side. America knows how many people it needs. The money, the reactors, and the policy are already moving. What it lacks is a workforce-development system preparing to scale at speed. The institutions that make nuclear technicians and operators are still built for a shrinking fleet. Every credential carries two to four years of lead time, so every month they wait is a month that cannot be bought back later.

That is what makes it urgent rather than merely difficult. The demand is set. The only variable left is whether the people who build the workforce start scaling now, fast enough and at enough volume to put trained crews in front of the reactors before the reactors wait on them.

03

What is pulling demand up

Four forces are arriving together, and none were priced into the 2024 forecast. They stack rather than trade off: the same technician is needed to restart old plants, build new ones, upgrade the fleet, and power the data centers paying for all of it.

Force 1 · Policy

The order to quadruple

Executive orders from May 2025 call for capacity to grow from 100 to 400 gigawatts by 2050. That means 10 large reactors under construction by 2030 and upgrades across the existing fleet.

≈ 300 new reactors · ~12 builds a year
Force 2 · Computing

AI's hunger for power

After twenty flat years, AI data centers are reshaping the grid. Tech firms have put more than $20 billion directly into nuclear to lock in steady, around-the-clock power.

$20B+ in private nuclear deals
Force 3 · Demographics

The retirement wave

About 40% of workers can retire within ten years. The mid-career people who should replace them were never hired. When seniors go, decades of plant knowledge go with them.

Losing knowledge, not just headcount
Force 4 · New plants

Small reactors and restarts

Small modular reactors and restarts like Palisades and Three Mile Island need full crews on tight schedules. The government has put more than $800 million into early small-reactor projects, and each new site starts staffing from zero.

Every new site needs a crew from scratch
04

Measuring the operator and technician gap

The best forward-looking model, from USC Marshall and the National Center for Energy Analytics, breaks the workforce into job types. Operators and technicians are the largest group, a fifth to a quarter of staff at a running plant, and the hardest to train.

Operators & technicians: projected demand
2025today
11,500–13,200 workers
2030near-term builds
13,400–15,800 workers
2050400 GW plan
grows with a 3–4× larger fleet
Today's workforce
What's needed
The one step that won't speed up

The model is blunt about where the wall is. By law, the Nuclear Regulatory Commission requires 18 to 24 months of training to license a single operator, and that timeline cannot be shortened. It is the one bottleneck that sets the pace for the entire build-out.

You can speed up money, reactors, and permits. You cannot speed up the making of a licensed operator below roughly two years. That makes training, not funding or regulation, the real limit on how fast America can add nuclear power.

05

How long it takes to train one person

There is no single timeline. Getting in the door is fast. Becoming a licensed operator is slow. To plan at scale, you have to plan each step on its own clock.

6–12 MONTHS

Certificate: the quick way in

A one-year community-college certificate, or an equivalent military credential, gets someone to an entry-level job. This is the fastest legitimate on-ramp.

2 YEARS

Associate degree: the standard credential

The two-year Associate of Applied Science in nuclear technology is the typical entry path. It splits into tracks like instrumentation, electrical, mechanical, and radiation protection. This is the workhorse degree.

+ 6 MONTHS – 2 YEARS

On-the-job training at a specific plant

Once hired, technicians spend six months to two years learning their plant's systems and safety rules under supervision. The learning never fully stops, because the technology keeps changing.

18–24 MONTHS · CANNOT BE SHORTENED

Reactor operator license

Becoming a licensed reactor operator means clearing a federal training and licensing process that cannot be rushed. This is the gate that paces the whole build-out.

~12mo
Fastest way in, by certificate or military credential
NUCP / DoD
~2yr
Standard associate degree to a job
BLS / NUCP
18–24mo
Operator license, the step that can't be shortened
USC / NCEA
3–5yr
Raw recruit to fully licensed operator
Composite
06

Who trains this workforce today

Three pipelines feed the industry. The first is a network of about 36 community colleges running a shared, industry-set curriculum. The second is the U.S. Navy, the largest trainer of nuclear operators in the world. The third is a thin layer of four-year degrees. The catch: the college network is built to size each class to the jobs available, on purpose.

InstitutionProgramTypeNote
U.S. Navy, NNPTC
Goose Creek, SC
Nuclear Power School + prototypeMilitaryThe world's largest source of nuclear operators, and the main feeder of cleared, trained talent into commercial plants.
Bismarck State College
North Dakota
Nuclear Power TechnologyAAS BASA founding college of the shared curriculum, with online and on-campus tracks.
Columbia Basin College
Washington
Radiation Protection / Operator / I&CAAS CertPartnered with Energy Northwest and the Hanford site. Offers both a two-year degree and shorter certificates.
Wharton County Junior College
Texas
Nuclear Power TechnologyAAS CertFour specialty tracks, tied to the South Texas Project plant and Texas A&M.
Cape Fear Community College
North Carolina
Nuclear Technology: reactor field techAAS CertHands-on learning at a working plant, plus high-school and Navy transfer routes.
Augusta Technical College
Georgia
Nuclear Engineering TechnologyAASA 64-credit program near the Vogtle plant, training techs and non-licensed operators.
Excelsior UniversityNuclear Engineering TechnologyBSThe largest source of four-year nuclear degrees, built for working adults and veterans.
Idaho State / Arkansas TechNuclear / energy technologyAAS BSUniversity programs located near national labs and operating plants.
The flaw in today's system

The college network is built to match supply to demand. Colleges report how many graduates the industry needs, and they train to that number, on purpose. That worked in a flat market. In a booming one it backfires: the system is designed not to overproduce at the exact moment overproduction is the goal. Class sizes are tiny, often a handful to a few dozen per program. The Navy produces scale, but on its own schedule, not the commercial fleet's.

07

How to fix it at scale

If the limit is a two-year training pipeline, the whole problem becomes one of throughput. Widen the entrance, shorten every step that can be shortened, and bring in people who already clear the hardest gates. A serious plan runs all five of these at once.

01

Recruit Navy veterans and adjacent workers first

The fastest way to qualified technicians is to hire people who already have most of the training. Navy nuclear veterans, and operators from oil, gas, chemicals, and power, clear the security, math, and discipline hurdles that take civilians years. They have the aptitude. What's missing is a clear path that credits what they already know.

Payoff: productive in months, not years
02

Break the degree into employable steps

Don't make every recruit finish a two-year degree before they're useful. Start with a 6-to-12-month certificate that produces a working technician, then let them earn the full degree and license while on the job and getting paid. This fills maintenance and outage roles fast while the operator pipeline runs in parallel.

Payoff: earlier pay, fewer dropouts, faster hiring
03

Build the school into the plant

Pair every new build, restart, and small reactor with a training program on site, the way a teaching hospital trains doctors on real patients. Students do supervised work on the actual plant, which folds on-the-job training into the degree and guarantees a job at the end. New sites staff from zero today; they should staff from a school built into the site.

Payoff: training and hiring become one motion
04

Stop sizing classes to today's jobs

The college network's match-supply-to-demand rule has to be flipped. Programs should be funded to train for the 2050 fleet, not the current one, backed by federal workforce grants like the $49.7 million awarded to 10 universities in April 2026. The risk of a few extra $100,000-a-year technicians is trivial next to the risk of a reactor sitting idle for lack of crew.

Payoff: removes the cap on supply
05

Grow fast without lowering the bar

Scale cannot come at the cost of safety, which is what makes nuclear work at all. Hold quality with the shared national curriculum, heavy simulator time, and clear data on every trainee's progress. The two-year licensing gate stays in place; it's a safeguard, not waste. Everything around it is what gets faster.

Payoff: more people, same safety standard
08

The bottom line

Where this leaves us

The verdict is bad, and the official numbers hide it. They forecast a fleet that was shrinking, not the one the country has now committed to build. Demand is set to nearly quadruple the workforce by 2050 while 40% of today's workers retire, and the training system produces only a fraction of what's needed, with two to four years of lead time on every recruit.

The limit is not money, reactors, or permits. Those are already moving. The limit is the two-year training pipeline, sitting on top of a school system built to never overproduce. That is the real bottleneck on American nuclear power.

Start now
The mandate
Timing is the whole game.

To deliver on the federal orders and the private deals already signed, the workforce has to be built ahead of the curve, not behind it. A program that opens in 2026 first graduates people around 2028 to 2030, right when the new plants come online and the retirements peak. America has done hard technical training at scale before; the Navy nuclear program proves it. The fixes are known and partly funded. What's missing is speed.

Whoever recruits veterans fast, turns the degree into paid steps, builds the school into the plant, and lifts the supply cap without lowering the safety bar will decide how fast America can power its future.

About Project Arklight

Project Arklight is a workforce-development company rebuilding how America trains skilled industrial labor.

We run a software-enabled trade school, Trade School 2.0, that assesses, trains, and deploys production-ready operators (electricians, machinists, welders, fabricators) to the companies reshoring American manufacturing and energy. We also publish original research on the skilled-labor gap: where it is, how deep it runs, and what it takes to close it. A shortage of skilled workers is the biggest obstacle to rebuilding American industry. As the nuclear build-out shows, the limit is the supply of trained people. Project Arklight exists to remove it.

Trade School 2.0

We build the builders.
Build your workforce.

You can't scale what you can't staff. We assess, train, and deploy operators who ship to spec on day one, co-designed to your floor and paid on outcomes. Co-design your first cohort.