Why hiring hasn’t matched output

Output continued to climb last year, but growth decelerated: the economy expanded at an annual rate close to 2.2% in 2025, down from about 2.8% the prior year. Firms simultaneously poured capital into artificial intelligence projects — buying equipment, funding R&D, and building data centers — which lifted measured output even as new payroll creation remained modest.

The jobs picture is muted. Aggregate payroll additions were tiny relative to past cycles, roughly 181,000 positions for the year, while the headline unemployment rate hovered near 4.3%. Monthly data are noisy: January’s nonfarm payrolls rose by about 130,000, but revisions trimmed prior months and methodological changes to the BLS birth‑and‑death adjustment and pending household population-control updates mean single-month readings should be interpreted cautiously.

At the same time, a deeper reconfiguration is visible beneath the headline: roughly 1.8 million Americans have been jobless for six months or longer — about one in four of the unemployed — pointing to weaker hiring rather than elevated separations. Corporate trackers logged more than 1.2 million job‑cut notifications across 2025, and planned layoff announcements in January exceeded 108,435, even as planned hires fell to only 5,306 in that month.

Two forces are intersecting. First, rapid AI deployment and outsized capital commitments are substituting capital for some routine labor and redirecting income toward shareholders and higher‑income households. Industry estimates put global AI infrastructure spending near $1.5 trillion in 2025; concentrated procurement through a small set of hyperscalers amplifies capital intensity and vendor lock‑in, while delays and permitting frictions have already reshaped about $64 billion of planned U.S. data‑center projects.

Second, immigration flows that historically supplied many workers have fallen sharply — forecasts point to net inflows near 160,000 versus roughly 1.1 million in typical years — shrinking the pool of available workers and eroding the local consumer demand that encourages hiring in certain regions and sectors.

• Researchers estimate AI‑related spending added about 0.4 percentage points to GDP last year, but much of the spending is capital‑intensive and accrues to owners of capital rather than wage earners.
• A research strand warns roughly $1 trillion of current consumption is supported by asset appreciation and capital flows rather than broad‑based wage growth, a gap that could reverse if investment expectations falter.
• Entry‑level opportunities have been squeezed: postings for early‑career roles have declined materially (entry‑level listings down roughly 35% versus early 2023), lengthening hiring cycles and increasing applicant competition for each opening.

The result is a growth pattern sometimes described as an economy that’s expanding with only modest job creation — growth that benefits capital owners faster than many workers. Sectoral outcomes diverge: manufacturing and agriculture report persistent shortages tied to reduced foreign‑born employment and higher input or trade‑policy costs, while many tech employers shift toward a smaller set of durable, AI‑capable roles even as they announce high‑profile layoffs in other divisions.

Corporate behavior has shifted toward a “low‑hire, low‑fire” posture in many firms — preserving payrolls but throttling new entry and lateral movement — which can mask micro‑level slack and raise long‑term unemployment risks. For jobseekers, longer searches and reduced early‑career openings translate into pay compression and scarring: rehired workers often accept pay cuts and prolonged unemployment can erode skills.

Policymakers face a trade‑off. Sustained capital investment in new technologies can raise long‑run productivity, yet transitions without scaled retraining, portable credentials, streamlined immigration for shortage occupations, and measures to reduce vendor concentration risk persistent dislocation and uneven regional outcomes. Watch for debates over targeted hiring supports, immigration channels tied to sectoral needs, competition policy for hyperscalers, and fiscal measures to bolster wage‑backed demand.