24AUG

Jobs open, hires fall, training cut

3 min read

16:17UTC

US job openings jumped to 7.6 million in April while actual hires fell to 5.1 million, the signature of a labour market where employers want workers they cannot find.

← Back to Four methods, one answer on AI and jobs Jump to analysis ↓

EconomicDeveloping

Key takeaway

US openings hit 7.6 million while hires fell to 5.1 million, the mark of a skills mismatch.

US job openings jumped to 7.6 million in April, the highest since May 2024, while actual hires fell to 5.1 million, the Bureau of Labor Statistics reported in its Job Openings and Labour Turnover Survey ¹. The survey, known as JOLTS, tracks how many posts employers advertise against how many they actually fill. Posts sitting open and unfilled are what a skills mismatch looks like when employers rewrite what they want faster than the labour pool can follow.

LinkedIn's 2026 workplace learning data shows the share of firms offering AI upskilling fell from 35% to 26% year-on-year ², so the training that would close that mismatch is shrinking just as the gap widens. Companies are pushing workers to adapt to AI while cutting the instruction that would let them, even as capital pours into data centres. Cut staff for AI, spend on AI, teach the remaining workers less about it.

Fed Governor Lisa Cook named the displacement risk directly in late May, citing widening software-sector bond spreads . Bond spreads widen when investors demand more to lend to a sector they judge riskier, so a central banker pointing at software borrowing costs is reading the same signal the openings data shows: the corner of the economy most exposed to AI is the one where hiring has stalled and credit has tightened first.

Deep Analysis

In plain English

In April 2026, US employers posted 7.6 million job openings; the highest since early 2024. Actual hires fell to 5.1 million in the same month. The gap exists because companies are rewriting job descriptions to require AI skills that most applicants do not yet have, then waiting. LinkedIn's 2026 data shows the share of firms offering formal AI upskilling fell from 35% to 26% in one year. Employers are raising the bar, cutting the ladder, and blaming the resulting vacancies on a skills shortage that they are partly creating.

Deep Analysis

Root Causes

Two structural forces explain the April 2026 mismatch.

First, AI tools have accelerated the rate at which firms can rewrite job descriptions to incorporate new technical requirements; effectively raising the floor faster than the training market responds.

LinkedIn's finding that firms cut formal AI training by 9 percentage points (35% to 26%) while simultaneously rewriting job ads to require AI literacy is contradictory only if firms intend to close the gap through hiring; it is coherent if firms intend to let the gap filter out candidates who have not self-trained.

Second, Fed Governor Lisa Cook's specific framing; widening software-sector bond spreads; points to a credit-market feedback loop. As AI compresses expected revenue growth in software-as-a-service, bond investors demand higher yields, which raises financing costs, which causes firms to stretch headcount further to preserve margins. The 7.6 million to 5.1 million openings-to-hires gap reflects that dynamic: posted jobs are aspirational, not budgeted.

What could happen next?

Consequence
If formal AI training continues to fall while job requirements rise, the skills gap becomes self-reinforcing: only workers who can afford self-funded upskilling clear the hiring bar, concentrating AI-era employment gains in higher-income cohorts.
Risk
Fed Governor Cook's citation of widening software-sector bond spreads suggests credit markets are already pricing an AI-driven revenue contraction in software, which would further suppress hiring budgets in the sector most responsible for AI adoption.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: The 1990s IT revolution produced a structurally identical gap: job openings in technology and finance surged from 1995 to 2000, but hires lagged as employers rewrote role requirements faster than training institutions could certify workers.

The National Center on Education and the Economy documented in 2001 that the gap peaked in 1999-2000; just before the dot-com crash; and closed only when employers lowered their stated requirements rather than when workers upgraded their skills.

Counter-parallel: Germany's Kurzarbeit (short-time working) scheme in 2008-2010 provides a case where a skills mismatch was largely avoided: rather than posting new requirements and leaving roles vacant, employers kept workers on reduced hours while retraining. The result was that German unemployment peaked at 5.3% versus the US 10% during the same recession. No equivalent mechanism exists in the current US or UK policy environment.

The openings-to-hires ratio (7.6 million to 5.1 million) is now running at 1.49, significantly above the 1.2 ratio that prevailed in 2019. That elevated ratio historically signals either an unusually tight labour market or a structural mismatch. Given that UK vacancies are falling (707,000, the lowest since early 2021) while US openings are rising, the likely explanation is divergence: US firms are re-specifying roles upward while UK firms are withdrawing postings entirely as demand softens.

Fed Governor Cook's warning about software-sector bond spreads is the underpriced risk in this picture. Software company high-yield spreads widening signals that fixed-income markets are pricing in AI-driven revenue compression in the sector, which will further tighten software hiring budgets over the next two to four quarters.

Consensus view: The MIT Sloan Management Review and Osterman Research have both documented the same pattern: when AI tools compress task time in a role, firms respond first by raising output expectations for remaining staff, then by quietly eliminating the role entirely at the next headcount review.

The April 2026 JOLTS data; 7.6 million openings but only 5.1 million actual hires; is consistent with firms posting roles they are not urgently trying to fill while they assess whether AI can cover the gap.

Counter-view: The New York Federal Reserve's Survey of Consumer Expectations shows that workers' own subjective job-finding probability held above its 2019 baseline through April 2026, suggesting that the labour market tightness is employer-side rather than a broad structural displacement. The NY Fed framing implies the JOLTS mismatch reflects elevated churn and role-rewriting rather than AI-driven elimination of positions.

Key tension: Whether the 9-percentage-point fall in formal AI training (LinkedIn: 35% to 26%) is firms becoming more selective about programme quality or firms withdrawing investment because they see AI as replacing rather than augmenting the roles they would otherwise train.

Sources:Bureau of Labor Statistics·LinkedIn

Mentions:Bureau of Labor Statistics →LinkedIn →Lisa Cook →Federal Reserve →Survey of Consumer Expectations →Germany →MIT Sloan School of Management →

First Reported In

Update #14 · The AI layoffs nobody is counting

Bureau of Labor Statistics· 20 Jun 2026

Read original →

Causes and effects

Caused by

Fed governor names AI job-loss risk

Fed Governor Cook's late May speech naming AI displacement risk is confirmed by the JOLTS and LinkedIn data showing an openings/hires bifurcation and declining upskilling.

Occurred 27 May 2026

Read story →

This Event

Jobs open, hires fall, training cut

An openings surge alongside falling hires is what a skills mismatch looks like, and it is precisely the channel a declared-layoff count cannot capture.

Different Perspectives

Office for National Statistics

Deferred its Transformed Labour Force Survey beyond November 2027 and disclosed a May 2026 telephone-collection failure. The ONS carries no AI-attribution layer at all, so Britain sits outside this month's cohort of measuring states by its own admission.

Uber India, Swiggy, Zomato and Urban Company

Named as respondents after the Karnataka High Court extended the interim welfare-fee deposit arrangement under the state's gig-worker welfare law to Uber India on 28 July, joining the other platforms already under the same order. The companies are contesting the underlying law while complying with the interim deposit terms.

Kenya State Department for ICT and the Digital Economy

Its draft AI policy, open for consultation to 4 August, proposes a pay floor for data-annotation work, where Kenyan annotators earn $1.46 to $3.74 an hour against $21 to $27 in the US, on figures relayed by the trade outlet WeeTracker. Kenya is legislating on AI labour even though the World Bank rates it among the least exposed economies.

ARAN and Italian public-sector unions

Signed the CCNL Funzioni Centrali 2025-2027 on 6 August, the first Italian national contract with a dedicated AI Title, barring fully automated employment decisions without meaningful human intervention and requiring advance union notice of AI deployment. The unions secured this through bargaining rather than waiting for legislation.

US employers reporting to Challenger, Gray & Christmas

Named artificial intelligence as the leading stated cause of job cuts for a fifth consecutive month in July, at 33% of that month's total, even as the overall cut count fell 27%. Employers kept citing AI as the reason even as scrutiny of the attribution rose.

Bank for International Settlements

Bulletin 130 reports a 0.75 percentage point average unemployment rise across high-AIPI countries between 2023 and 2025, while its own footnote 2 states the index is strongly correlated with employment shares in AI-exposed sectors it is used to predict. The bulletin calls the productivity payoff uncertain and uneven.