Most workforce-needs forecasts are wrong, and the way they are wrong matters. They overestimate growth in the headline-grabbing categories (data scientist, AI engineer, robotics technician) and dramatically underestimate growth in the categories that actually drive net new hiring (home health aide, RN, electrician, wind-turbine technician, solar installer). The Bureau of Labor Statistics' own 2024–2034 Employment Projections, the most authoritative U.S. forecast, is unambiguous on this. Of the ten occupations projected to add the most jobs over the next decade, only one — software developer — is what a layperson would call "tech." Five are in healthcare. Two are in construction and clean energy. The rest are in education and customer service.
This piece argues that the most useful frame for "future skills" is to stop treating the next decade as a discontinuity and start treating it as the most predictable labor market in recent memory. The big shifts — aging demographics, the energy transition, AI augmentation of knowledge work — are all visible from where we are standing. The skills that pay are the ones aligned with those shifts, and they are not particularly exotic.
What the data says about where jobs are coming from
The BLS Employment Projections are the only U.S. workforce forecast with a track record long enough to evaluate, and the track record is decent: 10-year occupational projections are typically within roughly 10–15% of actual outcomes for major categories. The current vintage projects the U.S. economy will add about 6.7 million jobs from 2024 to 2034, with growth concentrated in three categories.
First, healthcare and social assistance: roughly 2 million new jobs, more than any other supersector. Home health and personal care aides alone are projected to add about 820,000 positions, the largest single-occupation gain in the country. Driving this is demographics — the BLS notes that the U.S. population aged 65+ will grow from 18% to 22% of the total by 2034.
Second, the energy transition. The International Energy Agency's 2024 World Energy Employment report estimated global energy-sector jobs grew by 1.5 million in 2023 alone, with three-quarters of the gain in clean energy. In the U.S., the Inflation Reduction Act and Bipartisan Infrastructure Law have together unlocked north of $1 trillion in clean-energy and infrastructure investment, which the Department of Energy estimates will support roughly 1.5 million net new jobs by 2030. The Bureau projects solar installer and wind-turbine technician as the two fastest-growing occupations by percentage in the country.
Third, professional services with an AI overlay. The World Economic Forum's 2025 Future of Jobs Report identified data analyst, AI/ML specialist, and information security analyst as among the fastest-growing roles globally, but their projected absolute growth is smaller than most readers assume — together perhaps 250,000–400,000 net new U.S. jobs over the decade. These roles will continue to be high-wage, but the macro story is healthcare and energy, not AI hiring.
The skill clusters with the strongest evidence behind them
Healthcare and care-work credentials
The largest near-term workforce-development opportunity in the U.S. is expanding the supply of certified nursing assistants, medical assistants, registered nurses, and home health aides. The bottleneck is not demand — it is regulatory complexity, low pay floors, and a training pipeline that is undersized for the demographic shift coming. Per CMS data, the U.S. is short roughly 200,000–300,000 nurses by most estimates, with the shortage projected to grow. Per AARP's Caregiving in the U.S. 2024, the country also relies on roughly 53 million informal family caregivers, many of whom would professionalize into paid roles if the credentialing pathway were navigable. For more on this, our flagship analysis is in The Caregiver Workforce →.
The skilled trades
The Department of Labor's registered-apprentice base is still well below where it needs to be — about 640,000 active apprentices in 2024, versus more than 1.5 million in Germany on a far smaller population. Electricians, HVAC technicians, plumbers, welders, and equipment operators are all projected to grow faster than the labor force average, and median wages in these occupations have grown faster than the broader economy since 2020. The case for rebuilding the apprenticeship system is stronger now than at any time in the past 40 years; we make that case in detail at Apprenticeship 2.0 →.
Clean-energy installation and operations
Solar installers, wind-turbine technicians, battery-storage installers, EV-charging electricians, and grid-modernization workers are all projected to grow at double-digit percentages. The Brookings Institution's 2024 analysis estimated that meeting U.S. climate goals would require roughly 4 million additional clean-energy workers by 2035. Most of these jobs are physical, regulated, and AI-immune. The training pathways exist; the capacity does not.
AI fluency as a complement, not a specialization
The most reliable empirical finding from the past two years of AI-at-work research is that the gains accrue to workers who can use the tool, not to workers whose job is building it. Brynjolfsson, Li, and Raymond's 2023 Quarterly Journal of Economics study found 14% productivity gains for AI-assisted customer-support workers; the MIT/Wharton/BCG consulting study found 25–40% gains on appropriate tasks. The implication: a marketing manager who can prompt well is more valuable than one who can't, by margins that the labor market is already pricing in. A 6–10 hour applied training is the right investment for most knowledge workers.
Cybersecurity
The (ISC)² 2024 Cybersecurity Workforce Study estimates the global cybersecurity workforce gap at roughly 4 million unfilled positions, with about 470,000 of those in North America. This is one of the few "tech" categories where the absolute numbers are large and the credential-to-job pipeline is reasonably short. Mid-career transitions into cybersecurity remain among the highest-return reskilling pathways in the U.S. economy.
The forecasting mistakes worth avoiding
Two patterns make most workforce forecasts unreliable. First, conflating which technology is exciting with which skills are in demand. Generative AI is the most consequential technology of the decade; it is not, by itself, going to produce hundreds of thousands of net new AI-engineer jobs in the U.S. Most of its economic impact will be through augmentation of existing roles.
Second, ignoring physical and demographic constraints. The U.S. cannot replace 2 million home health aides with chatbots, because the work is physical and intimate; it cannot install gigawatts of solar without trained installers on roofs; it cannot expand semiconductor manufacturing without electricians and process technicians physically present in plants. McKinsey Global Institute's 2024 "A New Future of Work in America" analysis was clear on this point: the U.S. labor market's largest gaps are concentrated in physical, skilled, and care-intensive work, not in software.
The most under-forecast skills of the next decade are the ones a high-school student can train into in two years: home health, electrical, HVAC, solar installation, wind tech. The most over-forecast are the ones requiring a master's degree.
What individuals, employers, and policymakers should actually do
For individuals: take the BLS Employment Projections seriously, choose a growing occupation with a feasible training pathway, and treat AI fluency as an add-on to that core skill rather than as a career in itself. For employers: invest in apprenticeships and earn-while-you-learn programs that can be scaled fast; this is where the public-funding match is most generous and the talent pipeline most reliable. For policymakers: prioritize the credentialing-pathway improvements (interstate license portability for nurses, accelerated pathways for veterans into the trades, federally backed apprenticeship intermediaries) that are bottlenecks the market cannot fix on its own.
The forecast is, for once, not the hard part. The hard part is building the training capacity to meet a labor market we can already see.
Updated May 21, 2026. This piece was substantively rewritten as part of NWLB's 2026 editorial refresh.
