Gartner just confirmed what Klarna learned the hard way: 80% of companies cut people for AI, and none of it moved the needle.

The stock pops are gone, the bots are struggling, and the workers are getting rehired offshore at half the price.

The math always reconciles.

The only question is who signs the check.

Read. 👇

Sponsored by Career Intelligence Weekly

Most people follow the job market by watching job boards.

But the real signals are happening somewhere else.

Career Intelligence Weekly analyzes the latest business news, hiring trends, layoffs, and industry shifts through the lens of the hidden job market — the roles that are created, filled, or reshaped before they ever become public listings.

Each issue connects the dots between major headlines and what they actually mean for your career.

Where opportunities are quietly opening.
Which industries are shifting talent.
And what signals smart professionals should be paying attention to right now.

If you want to understand where the job market is really moving — not just where it's already been — this is a newsletter worth reading.

Subscribe to Career Intelligence Weekly and start seeing the hidden job market.

Klarna’s Receipt

Two years of AI-driven layoffs. The Gartner data just landed. The ROI isn't there.

Sebastian Siemiatkowski stood in front of cameras in 2024 and announced Klarna's chatbot was doing the work of 700 customer service agents. He said it like a victory.

Spring 2025. Siemiatkowski changed his tune. "We went too far."

The bot couldn't handle the hard calls. Customers wanted humans. Klarna started rehiring. Quietly. The new model: students, parents, rural workers logging in to clean up what the AI got wrong. Siemiatkowski himself called it "Uber-style."

That was the early warning. May 2026 turned it into the verdict.

80% cut. Zero ROI.

Gartner published the finding on May 11, 2026. Of 350 executives at billion-dollar-plus companies who had piloted AI, 80% reported workforce reductions. The cuts had no correlation with higher returns. The companies winning at AI weren't the ones firing people for it.

One sentence from a Gartner analyst, buried in the Fortune piece, said the loudest part. The layoffs "seem to us to be a kind of one-time exercise by many in small amounts, but not what translates to getting full ROI from their AI investment."

Translation: the cuts were theater. The returns came from somewhere else.

Wall Street saw it coming

Goldman Sachs analysts pulled the data late in 2025. Companies announcing layoffs that year, even when they dressed the cuts in AI efficiency language, saw their share prices fall an average of 2% against peers. Companies that called it "restructuring" got hit harder.

Ten years ago, a layoff popped the stock. Investors saw discipline. Now they see something else. They see panic in a hoodie.

Goldman's own language in the note: the market now reads layoff announcements as a "negative signal" about company prospects. The bounce is gone. The shorts are reading the tape.

The boomerang is here

Orgvue surveyed 1,163 senior business leaders across eight countries in early 2025. Of the 39% who had cut staff for AI, 55% admitted those redundancies were the wrong call. Another 34% watched additional employees quit after the AI rollout.

Forrester predicted in October 2025 that half of all AI-attributed layoffs would be reversed by 2027. Quietly. Offshore. Or on contract. Or at lower salaries.

Six months later, the prediction is on schedule. Bloomberg data shows roughly half of AI-attributed layoffs result in the same roles being rehired offshore. Amazon's Seattle job postings dropped from 22,700 in H1 2025 to 4,540 in Q1 2026. In the same quarter, Milan and Pisa became Amazon's second and third largest hiring cities. The work moved. The headcount line shrank. The wage bill changed continents.

Commonwealth Bank of Australia reversed 45 AI-driven layoffs in 2026 after concluding the roles were never redundant. Klarna did it first. IBM second. Commonwealth Bank made it three. The list is going to get longer.

The AI itself still isn't there

MIT's NANDA initiative studied 300 enterprise AI deployments and interviewed 150 leaders in mid-2025. 95% of generative AI pilots produced no measurable profit and loss impact. Five percent generated real returns.

The headline got argued over. The methodology got picked apart. But the directional finding held up across other studies, and Gartner's May 2026 data lands in the same place from a different angle.

The deeper number is uglier. METR, a non-profit AI research lab, ran a randomized controlled trial in 2025 with 16 experienced open-source developers. Half their tasks allowed AI tools. Half didn't. The developers using AI took 19% longer to complete their work.

Then METR asked the developers what they thought happened. The developers said they were 20% faster.

Sit with that for a second. A 39-point gap between what people felt and what people did. The same gap powering most boardroom layoff decisions.

Companies built their AI severance cases on that feeling.

The survivors cost more than the cuts saved

Stanford's Jeffrey Pfeffer has been hammering this one for forty years. Layoffs don't reliably cut costs. They don't increase productivity. They don't reliably increase share prices.

The data still backs him. A Leadership IQ survey found 74% of surviving employees reported a productivity decline after corporate layoffs. Two-thirds said they felt overworked. A third said they expected things to get worse.

The 2026 numbers make the human cost concrete. Tech sector unemployment hit 5.8% in early 2026, the highest level since the dot-com bust. Median time to re-employment for a laid-off tech worker climbed from 3.2 months in 2024 to 4.7 months. San Francisco office vacancy reached 36.7%. Seattle's sublease availability jumped 22% year over year.

That's the cost the spreadsheet doesn't show. The Slack message at 11pm asking if they should start applying. The senior engineer who knows the legacy code and decides to leave before the next round. The recruiter who can't fill the role you opened back up six months later because every candidate already heard the rumors.

A layoff that saves $5 million in salaries can cost $7 million in churn, slowdown, and contractor markup. The savings show up in the next quarter. The damage shows up in the next two years.

IBM ran a different play

IBM CEO Arvind Krishna told the Wall Street Journal in May 2025 that his company had replaced "a few hundred" HR workers with an AI agent called AskHR. The number got reported, repeated, and inflated. Viral posts claimed 8,000. The number was hundreds.

What Krishna also said, which got quoted less: IBM's total headcount went up. The AI moved the spend. The HR roles disappeared. Programmers, salespeople, and humans doing "critical thinking" work got hired in their place.

That's the model the market rewards. AI as a productivity multiplier when companies use it that way. AI as a justification for severance is what gets the 2% stock haircut and the Gartner zero-correlation finding.

One CEO redeployed savings into higher-value work. The other tried to put AI on the headcount line and ran into the wall Klarna hit a year earlier.

The 2026 wave

The Klarna-IBM-Commonwealth Bank lessons should have slowed the pattern. They didn't. They accelerated it.

Tech layoffs reached 85,411 year-to-date through April 2026, the sector's highest since 2023. AI was the top stated reason for two consecutive months. Block cut headcount nearly in half. Coinbase laid off 14% citing the need to be "AI-native." Meta begins eliminating 8,000 roles May 20. Atlassian cut 1,600 in March. GM cut 600 IT staff yesterday in an "AI skills swap," with sources confirming the company is rehiring across the same department for different titles.

Some of these companies will get the ROI they're promising. Most won't. Gartner's data already says so. The press releases are running on momentum the data has stopped supporting.

Three takeaways

For HR leaders. The Gartner finding from May 11 is the cleanest single line of ammunition you have ever had. 80% of companies cut for AI. Zero correlation with returns. The next time a chief financial officer floats AI-driven workforce cuts in a planning meeting, the data is now sitting on the table. The board does not want to be on the wrong side of that table. Bring the receipts.

For CEOs. The layoff announcement is no longer a free pass to a stock pop. Investors are reading the cuts as a tell. Gartner is reading them as theater. The upside is gone. The bill is coming. The CEOs who get ahead of this are the ones treating AI as a redeployment tool. The ones who don't are signing checks they can't see yet.

For job seekers. You aren't crazy. The market for the work you used to do is coming back. It might come back at lower pay, offshore, or under a different title. Or it might come back to you at higher rates as a contractor, after the company spent six months trying to do without you. Either way, the company that fired you for AI overestimated AI and underestimated you. Negotiate.

The receipt

Klarna's bot handled the volume. The brand bled. Siemiatkowski got there first because he could afford to admit it. IBM got there next. Commonwealth Bank made three. Gartner just made the verdict official.

Other companies are still pretending. They'll hire offshore, rebrand the new role, and hope nobody connects the dots. The dots are connecting on their own.

The math on AI layoffs always reconciles. The only question is who signs the check.

The HR Blotter

America Keeps Adding Jobs Men Won’t Take - The U.S. is still creating jobs, but many are landing in healthcare and caregiving fields where men remain scarce. Male employment has stalled as manufacturing, transportation, and warehousing shrink, while women gain ground through education, flexibility, and steadier growth sectors. The labor gap is narrowing less because women are catching up than because men keep stepping back.

ICE Finds A Visa Fraud Sewer - ICE says OPT has turned into a fraud factory, with foreign students listed at empty buildings, locked offices, crowded home addresses, and phantom employers. Investigators found alleged pay-to-stay schemes, fake worksites, offshore payroll claims, and companies that could not explain who their “employees” were. The crackdown is now moving from paperwork to door-knocking, and the fraud trail looks ugly.

California Raised Wages. Fast Food Cut Back - California forced fast-food wages to $20, and restaurants answered with thinner crews, shorter shifts, and more machines. A new study using cellphone data found an 8% staffing drop at affected chains, while economists still argue over how much pain the raise really caused. Higher pay landed, but the bill came due fast.

Layoffs Stay Low, Fed Stays Frozen - Jobless claims ticked up, but layoffs stayed low and the labor market kept its spine. Continuing claims fell to the lowest level since January 2024, giving the Fed little reason to rush rate cuts. Tech may be swinging the AI axe, but the broader jobs machine still refuses to crack.

Job Titles Are Losing Their Grip - Job titles are useful shortcuts, but paying people by title alone is lazy machinery dressed up as order. Skills-based pay can reward real value and widen opportunity, but only if companies can measure skills cleanly, audit bias, and avoid turning invisible work into invisible pay. The smart move is not killing titles; it is making them answer to skills.

…

The Jim Stroud Podcast

Not subscribed to The Jim Stroud Podcast? Then you’ve been flying blind. Here’s a sneak peek at the latest episode debuting tomorrow.

…

…

One more thing before I go…

Do you have a job that is hard to fill but, the hiring manager does not believe you? If so, you need an HFI report.

The Hiring Friction Index (HFI) report cuts through the noise by showing how hard a role will actually be to fill. It brings together talent availability, competitive pressure, and candidate visibility into a single score, helping teams quickly distinguish between a real talent shortage and a sourcing or access problem.

And guess what? Its free. Click here to try it out.

…