The Lens · Where The Work Will Be

Growth is 3.4% of the openings

Every “fastest-growing careers” list ranks the smallest part of the number. Across the whole US economy, genuinely new positions are 3.4% of the jobs that open each year. The other 96.6% is people leaving.

Dragonfly Lens · published 29 September 2026 · every figure carries the date it refers to and the date we checked it

A job opens for three different reasons, and only one of them is growth.

Somebody left the workforce. The US Bureau of Labor Statistics calls this a labour force exit — retirement is common, especially in older occupations, but BLS does not publish how much of this is retirement versus school, caregiving, disability or other reasons lasting four months or more.

Somebody moved to a different occupation. A cook becomes a delivery driver. The seat is empty; the person is still working.

The occupation genuinely grew. A position exists that did not exist before.

Between 2025 and 2035 the BLS projects about 17.5 million job openings a year across all US occupations.¹ Growth accounts for roughly 591,800 of them. That is 3.4%. Every headline built on “fastest growing” is ranking that 3.4% and ignoring the rest.

The short version

Solar photovoltaic installer is one of the fastest-growing occupations in America — 26.9% of its openings come from growth. It needs about 4,200 people a year. Carpentry needs about 62,800 — fifteen times as many people — even though growth is a much smaller share of its openings. If you want to know where the work is, growth is the wrong column.

Thirteen occupations, ranked by how many people they actually need

Occupation	Employed now	Openings a year	From growth	Share aged 55+	Median age
Home health and personal care aides	4,677,100	760,500	11.1%	32.7%	—
General and operations managers	3,599,000	285,000	6.4%	21.7%	44.0
Heavy and tractor-trailer truck drivers	2,221,200	214,500	3.9%	—	—
Nursing assistants	1,505,900	196,200	2.0%	19.9%	39.6
Registered nurses	3,465,400	180,800	10.8%	21.4%	42.3
Construction laborers	1,489,300	120,000	9.1%	18.2%	40.0
Software developers	1,717,800	95,300	18.3%	11.9%	38.6
Elementary school teachers	1,419,300	87,500	-0.7%	—	—
Electricians	821,000	72,700	10.4%	18.0%	39.6
Farmers and ranchers	788,700	70,800	-3.7%	54.4%	56.1
Carpenters	889,700	62,800	5.6%	20.5%	41.5
Plumbers, pipefitters and steamfitters	510,600	42,000	8.2%	18.4%	40.2
Solar photovoltaic installers	31,100	4,200	26.9%	—	—

Openings are annual averages over 2025–35, projected August 2026. “From growth” is new positions as a share of all openings — our arithmetic from the BLS columns.¹ Age figures come from a different BLS survey (the Current Population Survey) using a different occupation classification than the employment and openings columns. For some rows the two surveys' employment counts are close; for others they differ by up to 60%. Read the age figures as the shape of an occupation, not a precise match to the row's own employment number.²

These 13 were chosen to span healthcare, the trades, tech, management and agriculture — not ranked to make a point. They are not the 13 with the most openings economy-wide: fast food and counter workers alone need 827,400 people a year, more than any occupation in this table, including the one that leads it.¹

The clearest case is the one nobody calls a growth industry

Farming and ranching is shrinking. The BLS expects fewer of these jobs in 2035 than in 2025. It still needs about 70,800 people a year — more than carpenters need (62,800), just under what electricians need (72,700).

54.4% of American farmers and ranchers are 55 or older, and the median age is 56, with nearly a third already 65 or older.² Farmers do not separate from the occupation faster than the average worker overall — their total separation rate is close to the economy-wide figure. What differs is the mix: more of their separations are people leaving the workforce outright, and fewer are people transferring to a different occupation, which is exactly what an older workforce looks like.

The same pattern, in the trades. Carpentry's 62,800 annual openings break down as about 23,500 people leaving work altogether, 35,700 moving to a different occupation, and 3,500 genuinely new positions. Growth is the smallest piece by a wide margin — about a tenth of transfers, about a seventh of exits — and the trade still needs 62,800 people a year.

What this does not tell you

Four things, stated plainly, because a table like this invites over-reading.

It says nothing about pay. Home health and personal care aides top the table by a distance. Their median annual wage was $35,800 in May 2025.⁴ High demand and low pay coexist easily when the work is hard to automate and the buyer is a public budget.
It says nothing about how easy it is to get in. An occupation with many openings may also have many applicants, a licence requirement, or a training place that takes years. That is the next thing we are building.
Transfers cut both ways. People leave carpentry for other occupations, and others arrive in carpentry from elsewhere. This shows seats emptying, not who fills them.
Projections are not promises. These come from a government agency with a published method³ that is explicitly built to reach a full-employment outcome in the target year — it does not forecast a recession or a boom along the way. BLS itself says the openings figure should not be read as a precise target, only as a general sense of scale. Automation is an assumption inside the model, not an observation. And BLS's own track record is uneven by sector: its most recent published evaluation found it beat a simple no-change baseline overall, but did noticeably worse specifically for healthcare support — the category this table leads with.¹⁰

One honest wrinkle in our own numbers. The 2025 age data excludes October 2025, because the government shutdown stopped collection that month.² It is an annual average of eleven months, not twelve. We do not think it moves anything here materially, and we would rather say it than have you find it.

Explain it to a friend

“Everyone ranks jobs by which ones are growing. But across the whole economy, new positions are only about 3% of the jobs that open every year — the rest is people leaving the workforce or moving on. American farming is shrinking and still needs seventy thousand people a year, because more than half of farmers are 55 or older and most of that replacement is people retiring outright, not switching careers. The question isn't which job is growing. It's who's about to leave.”

Check it yourself

Both tables are free, public and published by the BLS. The openings figures are table 1.10 of the Employment Projections programme,¹ released 27 August 2026. The age figures are table 11b of the Current Population Survey.² Our “from growth” column is the projected employment change divided by ten, divided by total annual openings. Nothing here required a paid data source.

A note for anyone comparing with Canada: the two countries do not define replacement the same way. The US counts people moving to another occupation as a separation; Canada's projection system counts only retirements, deaths and emigration, and treats occupation switchers as supply instead. The US replacement share is therefore structurally larger. They are not the same number and should not share a column — which is why this piece is US-only until we can do the Canadian side properly.

Sources

US Bureau of Labor Statistics, Employment Projections, Occupational separations and openings, projected 2025–35 (table 1.10), released 27 August 2026. Employment, projected employment, labour force exits, occupational transfers and total openings by occupation. — bls.gov
US Bureau of Labor Statistics, Current Population Survey, Employed persons by detailed occupation and age (table 11b), 2025 annual averages, last modified 20 February 2026. Excludes October 2025. — bls.gov
US Bureau of Labor Statistics, Occupational Separations Methodology and Employment Projections Data Definitions — method · definitions
US Bureau of Labor Statistics, Occupational Outlook Handbook, Home Health and Personal Care Aides: median annual wage $35,800 in May 2025. — bls.gov
US Bureau of Labor Statistics, Evaluating the 2014–24 Employment Projections, published 19 January 2024. Found BLS beat a no-change-in-shares baseline overall, but underperformed it specifically for the healthcare support occupational group. — bls.gov

BLS material is in the public domain. Figures retrieved 24 September 2026. Any errors in the arithmetic are ours, and corrections appear in our corrections log.

Dragonfly Lens publishes research, not advice. Nothing here is a recommendation about a career, an investment or anything else. Every figure names its source and the date the data refers to, so you can check it and disagree with us.