State of the Workforce July 2026 - The Loyalty Trap

Reading time: about 15 minutes. Jump to what the data means for your next move.

July 2026 produced one of the safest labor markets in 57 years and one of the least rewarding. Almost nobody was fired. Almost nobody moved. Payrolls actually shrank, and for the first time since 2022, the average American worker's pay bought less than it did a year ago. If you spent July feeling like you were standing still while the numbers on the news looked good, the numbers were not wrong and neither were you. They were measuring two different things.

This is the July edition of Metaintro's monthly State of the Workforce report. Last month we called it the AI alibi, companies naming AI as the reason for cuts that were really about cost. This month the story moved somewhere more uncomfortable. The cuts slowed down. The trap closed anyway.

The Loyalty Trap

Lacey Kaelani, CEO of Metaintro.

For twenty years the advice was simple. Find a good employer, do good work, stay long enough to compound. Loyalty was the safe play, and safe meant slow but sure.

July 2026 is the month that advice quietly stopped working.

Look at what the data actually says. You are less likely to be laid off than at any point since Nixon's first year in office. That is real, and if you have lived through a layoff you know it is not nothing. But look at what you get for that safety. Your pay is losing to inflation. Your odds of being hired somewhere else are the lowest in a decade. The people getting raises are the ones changing jobs, and changing jobs is the one thing this market will not let you do.

That is not security. That is a holding pattern with a nice name.

The cruelty of a low-hire, low-fire economy is that it looks like stability from the outside and feels like a slow leak from the inside. Nobody sends a memo when your real wages go negative. There is no announcement, no severance, no LinkedIn post. You simply notice, sometime around the third month, that the same paycheck covers less than it used to, and that the job listings you are qualified for either do not exist or pay what you already make.

Employers know this. A frozen market is a cheap market. When workers cannot leave, the pressure to pay them more disappears, and that shows up in the data almost immediately, posted wage growth is running about a full point below what current employees are getting, which means the cooling is not finished. It is queued.

So what do you actually do? Not panic, and not wait either. The workers who come out of a stuck market ahead are the ones who spend it getting more valuable rather than more comfortable, building a skill their employer would have to replace, making their work legible outside their own company, keeping a live network instead of a dormant one. The market will unfreeze. It always does. The question is whether you spend the freeze compounding or coasting.

Loyalty is not the mistake. Passive loyalty is.

What Did July's Jobs Report Show?

July's report broke June's pattern rather than extending it, and the way it broke matters more than the headline. Total nonfarm payroll employment fell by 23,000 in July, against economists' expectations of an 85,000 gain, and the unemployment rate edged down to 4.1%. Read no further and you would conclude the labor market finally cracked. Read one line further and you find the opposite. Private payrolls rose by 30,000. Government shed 53,000, and local government education alone accounted for 50,000 of that. The negative print is a public-payroll event, not a private-sector collapse, and almost every headline written on Friday morning obscured that distinction.

The revisions did the real damage. May's gain was cut by 66,000, from 129,000 to 63,000, and June's by 37,000, from +57,000 to +20,000. Together those two months are 103,000 jobs lower than previously reported. That reframes the first half of 2026 retroactively: the +57,000 June print this report described last month as a deceleration was, in fact, a near-standstill. Two of the last three months have now been revised down, which is the pattern to watch rather than any single release. Over the prior twelve months, nonfarm payrolls averaged a gain of just 34,000 a month.

On pay, the establishment survey confirmed the squeeze this report has been tracking all month. Average hourly earnings for all private employees rose 2 cents to $37.62 and were up 3.2% over the year. For production and nonsupervisory employees, average hourly earnings rose 4 cents to $32.40. The average workweek held at 34.3 hours. A 3.2% nominal gain is the same figure that, once adjusted for inflation, produced the first negative real-wage reading since 2022. Nothing in July's report changes that arithmetic. It confirms it.

The household survey carried the most uncomfortable number in the release, and it complicates the low-fire story. The number of people on temporary layoff increased by 153,000 to 921,000. That is not a rounding error, and it sits awkwardly beside a week of record-low jobless claims. The labor force participation rate held at 61.4%, but it has fallen 0.7 percentage points since January, and the employment-population ratio is down 0.5 points over the same stretch. There were 6.9 million unemployed people in July, of whom 1.8 million had been out of work for 27 weeks or more, accounting for 25.5% of the total.

The industry detail tracks the story this report has told all year, with one warning sign. Health care added 22,000 jobs and remained the only consistent engine, but that is well below its 36,000 average monthly gain over the prior year, which means even the engine is decelerating. Leisure and hospitality lost 40,000. Retail trade lost 19,000, concentrated in warehouse clubs and supercenters. Financial activities fell another 14,000 and is now down 121,000 from its May 2025 peak. Construction added 22,000, professional and business services 18,000, and information 11,000.

One useful cross-check. The ADP National Employment Report, published two days earlier, counted 44,000 private jobs added in July. The government's private-payroll figure came in at 30,000. Two independent measurements, built from different samples by different methods, landed 14,000 apart and pointed the same way. That agreement is worth more than either number alone, and it is the strongest evidence that private hiring in July was weakly positive rather than negative, whatever the headline said.

Why Is Almost Nobody Getting Laid Off?

Because by the clearest available measure, firing has nearly stopped.

In the week ending July 18, Americans filed 188,000 new claims for unemployment benefits, the lowest weekly figure since September 1969. That week was first reported at 187,000 and revised up by 1,000 in the Department of Labor's July 30 release, which is the figure used here. As Axios noted, the American labor force was far smaller in 1969, so measured against the number of people actually working today, this is the lowest level of jobless claims on record.

Two caveats matter, and most coverage skipped both.

First, this is a claims number, not a layoff-rate number. Those are different series, and they get conflated constantly. The federal layoffs rate, the share of workers separated involuntarily in a given month, held steady at 1.1% in June, per the latest JOLTS release. Low, but nowhere near a 57-year record.

Second, 188,000 was a single week. Claims rebounded to 197,000 the following week, up 9,000 from that revised level, with the four-week moving average at 202,750. July's real story is a low, gently rising plateau, not a historic trough.

Oxford Economics senior US economist Matthew Martin put the caveat plainly in a note quoted by Axios: "While seasonal factors may be impacting the headline number at the margins, the extremely low level of claims highlights a low layoff rate and the strength underlying the labor market."

So: genuinely, unusually safe. Just not quite as safe as the headline suggested.

So Why Does It Feel Like You're Falling Behind?

Because you are, and there is now a number for it.

Inflation-adjusted wages and salaries for private-sector workers fell 0.4% year over year in the second quarter of 2026, the first decline in workers' purchasing power since 2022. That figure comes from Indeed Hiring Lab economist Cory Stahle's July 31 analysis of the federal Employment Cost Index, adjusted for inflation.

Nominal pay did not collapse. It just stopped outrunning prices. Private-sector wage growth slowed to 3.1% year over year, down from 3.5% a year earlier, while inflation reaccelerated in the second quarter, driven in part, Stahle notes, by higher energy prices tied to the conflict in Iran. Slower raises plus faster prices equals less money, and that is the whole equation.

It gets worse on a lag. Indeed's Wage Tracker, which measures pay advertised in job postings rather than pay received by current employees, showed 2.4% growth in June, roughly 0.8 points below the government's least volatile wage series. Advertised pay reaches new hires before it reaches incumbents, so the Tracker tends to move first. Shift it forward about seven months and it maps closely onto the official measure. Translation: more cooling is already in the pipeline, and it has not arrived in your paycheck yet.

One genuinely interesting wrinkle. The gap between the fastest and slowest-paying industries is narrowing, down to 1.5 percentage points from 2.5 a year ago, with utilities leading at 4.1% and accommodation and food services trailing at 2.6%. The sectors cooling hardest are the ones that ran hottest after the pandemic: transportation and warehousing, real estate and leasing, and accommodation and food services each shed more than a full point of wage growth. The post-pandemic distortion is smoothing out. It is smoothing downward.

We covered what this means for individual raises in our breakdown of the real-wage decline and in why 2026 raises are getting smaller and more selective.

What Happened to Job-Hopping?

It stopped, which is precisely why the pay problem bites.

The quits rate held at 2.0% in June and the hires rate at 3.4%, according to Indeed Hiring Lab's Sneha Puri, writing on the June JOLTS data released August 4. Job openings were little changed at 7.4 million.

Quitting is how most workers get a raise. Decades of federal data show job switchers out-earn stayers, which is exactly what our reporting on the loyalty tax facing long-tenured workers found. When the quits rate falls to 2%, that mechanism switches off, not because workers stopped wanting more money, but because there is nowhere to go. We wrote about the shape of this in the low-hire, low-fire job market.

Puri's framing is the best one anybody offered in July: the labor market is a duck on a pond, calm on the surface and paddling hard underneath. The calm is real. So is the paddling.

There is also a supply-side story that deserves care. The civilian labor force has been shrinking since the end of 2025, which means some of what looks like weak hiring demand may be weak labor availability instead. Indeed attributes much of that decline to a falling foreign-born labor force. That is the finding as the data shows it; the policy argument around it is a separate conversation this report does not take a side in.

Who Cut Jobs in July, and What They Said

Cuts slowed in aggregate but did not stop, and July's list skewed large.

Three of these are worth pulling out.

UPS cut 48,000 jobs, leaned harder into automation, and then beat earnings. That combination is the one every worker in a physical-logistics role should be reading closely, because it is the clearest demonstration this year that headcount reduction and financial performance can move together in the short run.

Visa cut about 2,600 people, roughly 7% of staff, concentrated in tech and product, with the CEO memo pointing toward AI. We covered the Visa memo and what it told workers.

Microsoft cut 4,800, about 2.1% of its workforce, while spinning off four gaming studios, and separately offered laid-off staff up to 39 weeks of severance, an unusually generous package that says something about how tight specialised technical talent still is. Our full coverage is in Microsoft's 4,800 layoffs and the Xbox cuts, and the longer arc in how Microsoft's headcount fell across R&D roles.

Volkswagen belongs in July's story but not in July's ledger. The carmaker's restructuring, potentially up to 100,000 roles globally alongside four German plant closures, was first reported in late June, then worked through a July board meeting and CEO memo before the CFO addressed the plant closures and job losses at July earnings as profits sank. It is the largest restructuring in the company's history, and we tracked the internal fight over it in Volkswagen's boardroom budget battle.

The Cuts You Don't See

Announced layoffs are the visible layer. Underneath them, July surfaced a quieter mechanism: Inc. reported on the tactic of shrinking headcount without announcing anything, leaving roles unfilled, reorganising people out, letting attrition do the work. It never appears in a layoff tracker, which is precisely the appeal.

The reputational cost of doing it loudly is real and durable. Fast Company reported in July that the damage big layoffs do to a company's Glassdoor ratings can linger for years after the cuts stop, a strong incentive to shed people quietly instead.

And the pressure is landing on workers' expectations. More than half of workers told HR Dive's surveyed researchers they would consider accepting a lower offer after a layoff. In a market where leaving is already hard, that is how wage erosion spreads from the people who lose jobs to the people who keep them.

The sector concentration is stark. Challenger, Gray & Christmas found that tech accounted for nearly a third of all US layoffs in the first half of 2026, with sector job cuts up 83% year over year from January through June. Financial Times analysis published July 24 put the running total at roughly 140,000 tech jobs cut so far in 2026, even as the same companies pour record sums into AI infrastructure.

One July development deserves separate mention. Meta faces allegations that its AI-assisted layoff selection disproportionately flagged workers who had taken protected leave. Those are allegations in a filing, not established findings, and Meta disputes them. But with managers now openly reporting that they use AI to help make layoff decisions, it is the first significant legal test of who is accountable when the model picks.

Where Are the Jobs Actually Going?

Into healthcare, and into senior roles. Not much else.

Private education and health services added 68,000 hires in June, the largest gain of any sector. It is, at this point, close to the only reliable engine in the American labor market, a pattern our reporting has tracked through spreading nurse strikes and nursing schools recruiting more men to close the shortage. Even that engine has strain in it: clinic and physician practice bankruptcies are spiking.

Leisure and hospitality went the other way, hard. Hires fell 87,000 between May and June, a half-point drop in the sector's hires rate, and the steepest decline of any industry. Year over year, hospitality hiring is down 174,000, with construction second at 27,000.

One more number reframes the whole picture: businesses with fewer than 1,000 employees made 92% of May's hires, per Fast Company's July 29 reporting. The household-name employers dominate the layoff headlines. The actual hiring is happening at companies you have not heard of, which means a job search aimed only at the logos you recognise is aimed at 8% of the market.

The Ladder Lost Its Bottom Rungs

The single most consequential shift in July's data is where the openings are.

Entry-level job postings have been trending down since 2022 and, as of May, were down 7.5% from a year earlier, according to Indeed research cited by Axios. Senior-level postings over the same period were up nearly 15%.

That is not a hiring slowdown. That is a structural change in the shape of the job market, and it falls hardest on people trying to get in. We have tracked it through entry-level roles disappearing, PwC consolidating advisory hiring from 72 offices to 13, and the counter-case in IBM betting on entry-level workers while rivals cut. There is a real argument, laid out in why cutting entry-level jobs could backfire on employers, that this is a trap for companies too, you cannot promote a senior person you never hired as a junior.

Is the AI Layoff Story Starting to Break?

Something has cracked in it, and it is worth laying the evidence out precisely, because much of the coverage has garbled both the numbers and who they belong to.

The workforce analytics firm Orgvue, in annual research published in April 2025, found that 39% of business leaders said they had made employees redundant as a direct result of AI, and that among those leaders, 55% admit wrong decisions were made about those redundancies. A figure often cited alongside it belongs to a different study entirely: 32% of US hiring managers said they had eliminated a role primarily because of AI and later rehired for the same or a similar position, according to Robert Half data reported by CNBC in July 2026.

That framing matters, and a lot of coverage got it wrong. The widely repeated line "55% of leaders regret AI layoffs" overstates it, 55% is the share of the 39% who actually made AI-driven cuts, not a majority of all business leaders.

Even corrected, it is damning. A separate Orgvue study published in March 2026 found that 89% of senior leaders admit they rely on instinct when making key decisions, while only about a third consistently use data. In that same study, 63% of executives expected redundancies within six months, yet fewer than half of senior managers, 44%, knew those plans existed.

The supporting evidence stacked up all month. Nvidia CEO Jensen Huang argued in mid-July that the more AI his company uses, the more people it has to hire. An HR Dive-covered study found that adding talent, rather than cutting it, correlates with better revenue growth. And Inc.'s reading of a Microsoft internal memo made the case that AI can make work faster without fixing a broken process underneath it.

None of this means AI is not changing work. It plainly is, see SAP telling workers to invent their own AI roles or face cuts and AI being named in one of every three US job cuts in June. It means the specific claim that AI is currently replacing large numbers of workers is not holding up, and roughly a third of the companies that acted on it are hiring people back.

What Should You Actually Do in a Frozen Market?

The standard advice for a bad job market is to sit tight. In this one, sitting tight is the thing that costs you money. Here is what the July data actually implies.

Stop waiting for the market to reopen before you prepare. The quits rate has been at or below 2% for months and the wage data says more cooling is queued, not finished. If your plan is to wait for conditions to improve and then move, you are planning to compete with everyone else who had that plan. The preparation (portfolio, references, a current sense of what you are worth) is the part you can do while frozen.

Aim at the 92%. Small and mid-sized employers made the overwhelming majority of hires. If your search consists of the twenty companies whose names you already know, you are fishing in the smallest part of the pond, and it is the part with the most competition.

Treat a lateral move as a real option. With senior postings up about 15% and entry-level down 7.5%, seniority is where the openings are. Moving sideways into a role that is one rung further along, at a smaller company, is currently a more available path than moving straight up where you are.

Know your number before the conversation. Real pay is falling, which means a flat raise is a pay cut and a 2% raise is close to one. Going into a review without knowing the market rate for your role is how the loyalty tax gets collected quietly. Our guide to what smaller, more selective 2026 raises mean covers how to frame that conversation.

If you are early-career, optimise for the rung, not the logo. The entry-level squeeze is structural, not seasonal. A role at an unglamorous employer that actually teaches you something beats a prestigious job queue that never opens. IBM's bet on entry-level workers while rivals cut is the shape of employer worth targeting right now.

And if you were laid off: the market is genuinely bad for speed and genuinely fine for eventual placement. More than half of workers said they would take a lower offer after a layoff, understandable, and worth resisting for a few weeks longer than feels comfortable, because the first offer in a frozen market is usually the employer testing how frozen you are.

The Bigger Picture

Put July's numbers side by side and the shape is unmistakable.

Firing is at a generational low. Hiring is at a decade low. Quitting is at a decade low. Real pay is negative for the first time in four years. The labor force is shrinking. Entry-level opportunity is contracting while senior opportunity expands. And roughly a third of employers who cut on an AI thesis are quietly reversing.

This is not a recession. A recession fires people, and this market conspicuously does not. It is something less familiar: an economy that has stopped moving people. Not a crash, a clot.

For workers, the practical consequence is that the two reliable tools for getting ahead (leaving for more money, or being recruited on the strength of a strong market) are both currently unavailable. What remains is the slower route: becoming demonstrably harder to replace.

Three Things to Expect in Q3

Wage growth keeps cooling before it recovers. Indeed's posted-wage data leads the official measure by roughly seven months and is running about 0.8 points below it. That gap is not a forecast, it is already-collected data that has not reached incumbent paychecks yet.

The AI-layoff reversal gets louder. With about a third of US hiring managers reporting they have already rehired for an AI-eliminated role, and 55% of the Orgvue leaders who cut on an AI thesis calling the decision wrong, the rehiring story moves from trade press into mainstream coverage, and probably into a few earnings calls.

Healthcare stays the engine, and gets tighter. Private education and health services is carrying the hiring numbers almost alone, against a shrinking labor force and rising practice bankruptcies. Expect harder competition for the same clinical staff.

What to Watch in August

• Whether claims stay near 200,000. The four-week average at 202,750 is the number to track, not any single week. A sustained move above 220,000 would be the first real crack in the low-fire story.
• The August jobs report, Friday September 4, and whether July's print holds up after revision. Two of the last three months have been revised down.
• July JOLTS, due early September. Specifically the quits rate. A move back above 2.2% would be the earliest signal that workers can move again.
• Rehiring disclosures. With 32% of US hiring managers reporting they have already rehired for a role they cut because of AI, watch for companies quietly reversing course, it tends to surface in earnings calls before press releases.
• Entry-level postings. If the 7.5% annual decline deepens, the ladder problem becomes structural rather than cyclical.

Methodology

This report draws on federal labor data, specifically the Employment Situation for July 2026 released August 7, the Employment Cost Index, the Job Openings and Labor Turnover Survey, and Department of Labor weekly unemployment insurance claims. Payroll, earnings, unemployment and participation figures are taken directly from the Bureau of Labor Statistics release and its data tables. Employment Cost Index and JOLTS figures are as interpreted by Indeed Hiring Lab economists Cory Stahle and Sneha Puri. The report also uses company disclosures, SEC filings and contemporaneous reporting from July 1 to 31, 2026.

Inflation-adjusted wage figures are Indeed Hiring Lab calculations combining the Employment Cost Index with consumer price data, not direct federal publications. Layoff figures are announced totals, which differ from completed separations. Job-postings data reflects Indeed's index and covers advertised roles rather than all hiring. Where a widely circulated figure was found to misstate its own source (the "57-year low" and the "55% of leaders" statistics both did), the corrected version is used and the discrepancy is noted in the text.

Every statistic here is attributed to its originating organisation with its collection period stated. Nothing is estimated, modelled or forecast by Metaintro.

Sources

Government Data and Official Releases

• Job Openings and Labor Turnover Survey, June 2026. US Bureau of Labor Statistics, released August 4, 2026
• Employment Cost Index, Q2 2026. US Bureau of Labor Statistics, released July 31, 2026
• Unemployment Insurance Weekly Claims. US Department of Labor, weeks ending July 18 and July 25, 2026, released July 30, 2026
• The Employment Situation, July 2026 (USDL-26-1291), US Bureau of Labor Statistics, released August 7, 2026

Labor Market Research

• June 2026 JOLTS Report: The Labor Market is a Duck on a Pond. Sneha Puri, Indeed Hiring Lab, August 4, 2026
• Q2 2026 Employment Cost Index: Losing Ground. Cory Stahle, Indeed Hiring Lab, July 31, 2026
• US Labor Market Snapshot, June 2026. Indeed Hiring Lab, July 23, 2026
• 55% of businesses admit wrong decisions in making employees redundant when bringing AI into the workforce. Orgvue, April 29, 2025
• Unintentional decision-making sets up AI deployments to fail. Orgvue, March 25, 2026

News and Press Coverage

• Low layoff rates defy AI jobs domination fears. Axios, July 23, 2026
• Employers who laid off workers citing AI are already starting to regret it. CNBC, July 1, 2026, reporting Robert Half data
• Layoffs just hit a 57-year low. So why does the job market feel so tough?. Fast Company, July 23, 2026
• US tech groups cut 140,000 jobs despite AI spending boom. Financial Times, July 24, 2026
• Tech layoffs in the first half of 2026. HR Dive on Challenger, Gray & Christmas, July 2, 2026
• Managers say they are using AI to make layoff decisions. HR Dive, July 31, 2026
• AI can make work faster, but it can't fix a broken process. Inc., July 18, 2026
• Adding talent, not cutting it, leads to better revenue growth. HR Dive, July 16, 2026

Recommended Reading

• Real Wages Fell in 2026, What It Means for Your Raise
• The Loyalty Tax Facing Long-Tenured Workers
• The Low-Hire, Low-Fire Job Market
• Entry-Level Roles Are Disappearing
• Why Cutting Entry-Level Jobs Could Backfire on Employers
• Visa's 2,600 Job Cuts and the Layoff Memo
• Microsoft's 4,800 Layoffs and the Xbox Cuts
• BuzzFeed Cuts 180 Jobs as Byron Allen Moves to Stem Losses
• 1 in 4 Workers Are Stuck in Jobs for the Health Insurance
• State of the Workforce June 2026, The AI Alibi

Stay ahead of the market. Metaintro tracks the labor market daily so you do not have to. Join free and get the roles, the pay data and the market signal in one place.