On August 27, 2026, the Bureau of Labor Statistics released its Employment Projections for 2025-2035: the government's official 10-year forecast of where American jobs are headed. It is the single most data-dense roadmap the U.S. labor market produces, and most recruiters will read a three-paragraph summary and move on.
That is a mistake. The projections tell you not just which sectors are growing but which occupations to build pipelines in, which clients are going to be the hardest to hire for in three years, and which categories of talent to stop chasing because the demand is structurally declining. Here is the complete read.
The Headline Number Is Not the Story
Total U.S. employment is projected to grow from 170.3 million in 2025 to 176.2 million in 2035, a net gain of 5.9 million jobs or 3.5% over the full decade, according to BLS.
On its face, that sounds like steady growth. Put it in context: 5.9 million jobs over 10 years is roughly 49,000 net new positions per month. The ADP National Employment Report for August 2026 counted 54,000 private sector additions in a single month that was considered a significant miss. A 10-year growth projection that only modestly exceeds one recent monthly print tells you this is a slow-growth economy with deep structural shifts underneath.
The aggregate number also obscures the most important feature of the forecast: the distribution is not remotely uniform. One sector accounts for 37% of all projected net gains. One occupational group accounts for the largest single block of projected losses. Everything else fills in around them. The recruiters who understand both ends of that distribution will make very different decisions than those running on aggregate optimism.
Healthcare Takes 37% of Every New Job
BLS projects the healthcare and social assistance sector to add more than 2.2 million jobs from 2025 to 2035, representing roughly 37% of all net new employment over the decade. That is a 9.5% sector-level growth rate, nearly triple the overall economy's 3.5% pace.
Within that sector, the occupational data is even more specific.
Healthcare support occupations (a major group covering home health aides, nursing assistants, medical assistants, and related roles) are projected to grow 13.3%, the fastest growth rate among all 22 major occupational groups BLS tracks. Healthcare practitioners and technical occupations (the group that includes RNs, nurse practitioners, and diagnostic technicians) are projected to grow 8.0%.
At the individual occupation level, nurse practitioners are projected to grow 41.0% between 2025 and 2035, the fastest rate of any single occupation in the BLS dataset. Home health and personal care aides are projected to add 847,300 jobs, the single largest numerical gain of any occupation in the country.
For recruiters who specialize in healthcare, this should function less as a data point and more as a directional mandate. The structural tailwind running through your market is not a temporary conditions story. It is a demographic arithmetic story: 73 million Baby Boomers aging into the period of peak healthcare utilization over the same 10-year window BLS is projecting. That demand does not reverse. It compounds.
The practical implication: pipeline built now in nursing, home health, nurse practitioners, physical therapy, and occupational therapy is pipeline built into a market that will be more competitive, not less, in 2028 and 2030. The recruiters who are signing long-term search agreements with healthcare systems today are locking in client relationships for a supply-shortage decade.
AI Gets the Credit, and the Body Count, in Office and Admin
The largest projected numerical employment loss in the BLS data is in office and administrative support occupations: a decline of 4.0% from 2025 to 2035, erasing 752,100 positions.
BLS is explicit about the cause. The official summary cites "the continuing integration of automation tools, including AI-powered systems, into workplace processes" as the primary driver. This is the federal government's labor statistics arm confirming, in writing, that AI is eliminating office and administrative jobs at scale. Not eventually, not theoretically.
Sales and related occupations are projected to decline 1.4%. Production occupations are projected to decline 0.4%. Retail trade loses an estimated 27,500 jobs over the decade as e-commerce reduces demand for physical retail employment.
For recruiters who work in administrative staffing, executive assistant placement, or sales support roles, this is not a "hedge your bets" data point. It is a structural market contraction forecast from the primary source. The market for those roles will be smaller in 2030 than it is today. The clients who currently hire heavily for administrative functions are going to hire less. Not because their businesses are struggling, but because software is doing the work.
The correct response is not to argue with the forecast. It is to shift capacity toward the sectors that are growing before the contraction arrives.
Tech Recruiting Is a Skills Story, Not a Headcount Story
Computer and mathematical occupations are projected to grow 7.3% from 2025 to 2035, making them the fifth-fastest-growing major occupational group. Professional, scientific, and technical services (the sector where most of those jobs live) is projected to add 926,700 workers, making it the third-fastest-growing sector in the economy.
But the growth is not distributed evenly across technology roles. Data scientists are projected to grow 34.6%. Computer and information research scientists are projected to grow 21.8%. These are roles that design, train, and evaluate AI systems, not roles that maintain legacy systems, run QA cycles, or write routine application code.
That distinction matters because the segment of tech recruiting that has contracted hardest since 2022 is general software development. Indeed's August 2026 Hiring Lab data put software development postings at 74.4 on the Job Postings Index, 26% below the pre-pandemic baseline, with 71% of the recovery concentrated in senior and AI-adjacent roles. The BLS projections confirm that directional split will persist for the decade.
The median annual wage for computer and information technology occupations was $109,470 in May 2025, according to BLS, more than double the overall median of $50,980. That premium is not disappearing. It is concentrating in the roles that require the highest AI fluency. Recruiters who can reliably source data scientists and AI researchers are working in a high-value, scarce-supply corner of the market. Recruiters who built their tech books around general software development volume are in a structurally declining corner.
Construction Is the Sleeper Sector
The projections show construction and extraction occupations growing faster than the average for all occupations from 2025 to 2035. That is a significant finding, and it is underreported.
The mechanism is specific: CHIPS Act semiconductor fabrication facilities, data center construction tied to AI infrastructure buildout, grid modernization driven by electrification mandates, and continued reshoring of manufacturing capacity are all converting into multi-year construction hiring calendars. These projects are not short-cycle. A semiconductor fab takes three to four years to build. A data center campus takes 18 months minimum. The skilled trades demand generated by these projects is long-dated and not easily satisfied with foreign workers or quick retraining programs.
The median annual wage for construction and extraction occupations was $59,540 in May 2025, per BLS, or 17% above the overall median. For recruiters building practices in trades: electricians, pipefitters, ironworkers, and construction managers, the 10-year outlook is one of sustained demand in a chronically supply-constrained market.
Three Decisions the Projections Should Force Right Now
The BLS projections are a 10-year view, but the sourcing decisions they imply are not 10 years away. Pipelines built in 2026 support searches in 2027.
Reallocate toward healthcare or get priced out. Healthcare recruiting is going to get more competitive, not less. Every major staffing firm and RPO shop reads the same BLS data. The difference between a healthcare recruiting practice that commands client relationships in 2028 and one scrambling for spot placements is whether those relationships are being built today, not in response to the shortage.
Reframe tech recruiting around AI-adjacent roles. Data scientists at +34.6% and computer and information research scientists at +21.8% are growing at rates that rival nurse practitioners. The sourcing playbook for those roles (where they cluster, what their career trajectory looks like, how they evaluate opportunities) is different from the SWE playbook. Develop it now.
Talk to your administrative-heavy clients about what comes next. The clients who hire heavily in office and admin today are going to hire less of it. That conversation does not have to be a warning. It can be a repositioning: which roles in their organization are growing, where the AI-driven productivity gains will flow, and how to redirect recruiting resources toward the functions that will actually expand. That advisory conversation is worth more than a job order.
BlueLine's matching tools are built for the market the BLS data describes: supply-constrained sectors, specific skills, specialized sourcing. Start a free account at bluelinesearch.ai/register.