The US Bureau of Labor Statistics quietly published its ten-year employment map late last month, and Fortune’s read-through of it over the weekend deserves a career-compass lens. The headline numbers sound mundane — total US employment growing just 3.5 percent from 2025 to 2035 — but the detail underneath is the clearest official picture yet of what a decade of AI diffusion does to a labour market. It is not a story of collapse, and it is not a story of boom. It is a story of displacement within growth: whole sectors surging because of AI’s appetite for power, other occupations shrinking precisely because AI does their tasks.
The decade’s shape
The BLS projects total US employment rising 5.9 million to 176.2 million by 2035 — growth of 3.5 percent, a fraction of the 10.9 percent the prior decade managed. The economy is still adding jobs; it is simply adding far fewer, and where it adds them is the whole story.
Utilities grow fastest of all 20 sectors — 9.8 percent — because of AI. The BLS is explicit about the mechanism: nearly all the job growth comes from electric power generation, transmission and distribution, driven by rising electricity demand “including artificial intelligence power demands” from the data-centre build-out. Solar power generation employment is projected up 153 percent and wind up 62 percent within that. The AI boom’s biggest jobs story isn’t coders — it’s the people building and running the power grid that AI runs on.
Healthcare adds the most jobs — 2.2 million, 37 percent of all new US jobs. Growing 9.5 percent, second only to utilities, driven by an ageing population. Nurse practitioners up 41 percent, health services managers up 24 percent.
Professional, scientific and technical services — the sector most entangled with AI work — is the third-fastest grower at 8.6 percent, adding 926,700 jobs. Data scientists up 34.6 percent, computer and information research scientists up 21.8 percent. The decade’s fastest-growing industries are the four energy-generation ones — solar, wind, geothermal and other electric power — with solar PV installers (+42 percent) and wind turbine technicians (+50 percent) topping the occupational charts.
And the other side of the ledger: office and administrative support falls 4 percent — 752,100 jobs, the largest decline of any major occupational group — with the BLS naming AI-enabled productivity as a driver. Sales occupations drop 1.4 percent as e-commerce and AI-assisted sales processes absorb demand.
The honest reading
This is a projections exercise, not a forecast with a track record — the BLS itself stresses “the uncertainty about potential impacts remains very high” and re-issues annually precisely because AI’s trajectory is hard to model. The utilities figure also needs its context: 58,800 new utility jobs is a fast growth rate on a small base, while healthcare’s 2.2 million dwarfs it in absolute terms. And “administrative support shrinks” has been the automation story since the photocopier; the interesting question is whether this decade is genuinely different or just faster.
But the shape is consistent with everything we’ve tracked on this site: the Economist’s estimate that AI has created roughly a million US jobs against 200,000 AI-linked layoffs, Challenger data showing AI falling to fourth place as a layoff justification, and the AI wage premium hitting 62 percent. Official projections, private analyses and layoff trackers all triangulate the same direction: AI’s first-order labour-market effect in 2026 is churn and reallocation, not net destruction.
What it means for New Zealand
A 10-year US projection doesn’t map onto a 5.6-million-person economy — but three of its signals have direct NZ echoes.
The AI-power-demand signal is already live here. New Zealand’s data-centre pipeline is real, and the same physics applies: every MW of AI compute someone builds here needs lineies, electricians, and grid planners. The BLS’s “fastest-growing industry is electricity generation” line is one NZ can plausibly import, given both our electrification targets and data-centre interest. Electricians were already in shortage before AI turned up the demand curve.
The healthcare signal is NZ’s own demographic wave. Our ageing population mirrors the US driver — care demand is structural, not AI-dependent. If anything the AI angle is a relief valve: administrative automation in the health system is more likely to free nursing time than to replace care, which remains stubbornly hands-on.
The admin-support decline is the warning shot NZ should take most seriously. New Zealand’s public service has already been through significant restructuring, and administrative and contact-centre roles are among the most exposed to exactly the AI capabilities the BLS is modelling. If the US projection is even half right, the NZ policy question isn’t “will jobs disappear” but “where do the 752,100-equivalent workers go” — and whether reskilling pipelines exist for them. That’s a career question before it’s a government question.
FAQ
Which industries will grow fastest through 2035? Per the BLS 2025-35 projections, utilities grow fastest (9.8 percent, driven by AI and data-centre electricity demand), followed by healthcare and social assistance (9.5 percent, adding 2.2 million jobs) and professional/scientific/technical services (8.6 percent, adding 926,700).
Is AI reducing total employment? Not in the BLS projection — total US employment grows 5.9 million (3.5 percent) through 2035, though far slower than the prior decade’s 10.9 percent. The impact shows up in composition: AI-linked sectors like utilities grow strongly while office/admin support shrinks by 752,100 jobs.
What job is growing fastest in America? Wind turbine service technicians (50 percent growth projected) and solar photovoltaic installers (42 percent) top the BLS occupational list, with nurse practitioners (41 percent) and data scientists (34.6 percent) close behind.
Does this apply to New Zealand’s job market? The numbers don’t transfer directly — but the drivers do: NZ’s own data-centre build-out supports trades and grid roles, ageing demographics support healthcare demand, and admin-heavy roles face the same AI task-automation pressure here as in the US.
Related reading: The Economist Counts a Million New AI-Era Jobs in America, AI Just Lost Its Top Spot as a Layoff Excuse, and AI’s 4.8x Productivity Dividend.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
