In 1987, economist Robert Solow quipped: "You can see the computer age everywhere except in the productivity statistics." Thirty-nine years later, corporate America is reprising the same script — but this time the stakes are measured in half a million jobs.
A working paper published by the National Bureau of Economic Research, based on surveys of 750 U.S. chief financial officers, projects that AI-driven job cuts will reach 502,000 in 2026 — nine times the roughly 55,000 AI-attributed layoffs recorded in 2025. Companies from Oracle (30,000 cuts) to Meta (8,000) to Microsoft (8,750 flagged through a voluntary retirement program with notifications going out May 7) are restructuring in anticipation of an AI productivity dividend.
The problem: most of those companies can't actually find the dividend yet.
The Gap Between Expectation and Reality
The same body of research that identified the surge in AI-driven layoffs also surfaced a striking dissonance. When researchers surveyed thousands of CEOs on AI's impact, nearly 90% said AI had no measurable effect on employment or productivity over the past three years. Companies reported average AI-driven productivity gains of 1.8% in 2025 — already modest — but when researchers checked that figure against actual revenue and employment data, the real number came out significantly lower across every major industry sector.
A companion Fortune analysis of more than 350 public-company CEOs found that 66% plan to freeze or reduce hiring through the rest of 2026, citing AI as a primary rationale: wait for the technology to mature, let headcount drift down through attrition, redeploy the savings into model licenses and infrastructure. The logic sounds tidy until you notice that the same leaders are simultaneously telling researchers they've seen almost no return on the AI investment so far.
This is the Solow Paradox running live. Executives see the technology everywhere — in pitch decks, board presentations, every vendor conversation — and draw the reasonable-seeming conclusion that productivity must be rising. What the data shows is something more complicated: AI tools are being adopted but not yet deeply integrated into workflows in ways that compound across an organization. The gains are real but narrow, concentrated in specific tasks rather than structural output.
What the Labor Market Actually Looks Like Right Now
The macro backdrop makes the paradox sharper. According to the most recent Bureau of Labor Statistics data, the U.S. economy added 178,000 jobs in March 2026, with unemployment at 4.3% — elevated relative to the 2022–2023 lows but not catastrophic. Job openings came in at 6.9 million in February, per JOLTS, with hires falling to 4.8 million and the quits rate stuck at 2.0% for the seventh consecutive month. Workers aren't quitting because they don't have better options to run toward.
The Conference Board has taken to calling this a "low-hire, low-fire" equilibrium: companies aren't conducting mass layoffs, but they're also not filling open roles. The vacancy rate is falling faster than employment is. That gap — between jobs that exist on paper and jobs that are actually being filled — represents paralysis at the hiring-manager level, not a structural labor shortage.
Through May 1, 2026, there have been 155 layoff events affecting more than 100,000 workers. The daily average is roughly 830 job losses. The vast majority of those cuts are being attributed, at least in part, to AI displacement — even by companies that can't yet demonstrate where the AI productivity is going.
The Strategic Miscalculation
Here's the asymmetry that recruiters need to put in front of their leadership teams.
Companies freezing hiring are making a one-way bet: that AI will deliver productivity at scale before the competitive cost of understaffing accumulates. But this bet has been wrong before. During the late 1990s, companies that deferred technology investment in favor of watching the market — waiting for the ROI to become obvious — spent most of the following decade catching up. The ones that hired aggressively through uncertainty built the institutional knowledge and the teams that made later technology adoption stick.
AI adoption is following a similar curve. The productivity gains that researchers can't yet measure in aggregate are showing up granularly — in specific teams, in specific workflows, among employees who have been deliberately trained on these tools for 12 to 18 months. You don't get those gains by trimming headcount and hoping the software figures it out. You get them by hiring people who are ready to work alongside the tools, and giving them enough time to build the institutional muscle.
The companies freezing hiring right now are making a costly bet: that when the AI productivity arrives at scale, they'll be able to rebuild capacity quickly. History says they won't. Talent markets don't snap back like financial markets. Institutional knowledge disperses. Culture frays. The hiring ramp, when it finally comes, will be slower and more expensive than the freeze savings justified.
What Recruiters Should Do With This Data
Make the internal case. The NBER figure — 502,000 AI-driven layoffs projected against near-zero measured productivity gains — is a board-level argument for maintaining hiring capacity. If your leadership is citing "AI efficiency" as a reason to pause a search, this data directly challenges that premise. Request specifics: which workflows have shown measurable gains, in which teams, over what time horizon? The answer will usually be narrower than the hiring freeze suggests.
Target the displaced talent now. The same companies cutting the most workers — Oracle, Meta, Microsoft, Amazon — are releasing experienced engineers, product managers, and technical leads into a market where hiring is frozen elsewhere. KORE1's recent tracking of 2026 displaced tech talent finds that senior engineers with 8–15 years of experience and production cloud experience are being absorbed quickly by mid-market companies. This window is open, but it won't stay open: as the year progresses, that talent pool will settle.
Distinguish the freeze from the floor. Not all hiring freezes are the same. Some companies have genuinely paused all external hiring; others have frozen backfills while keeping strategic growth roles open. Knowing which category your clients or internal stakeholders are in determines your approach. Freezes that are budget-driven tend to lift when budget cycles reset. Freezes that are AI-theology-driven may persist longer but are also more likely to break suddenly when the competitive cost becomes visible.
Watch the May 5 JOLTS release. The Bureau of Labor Statistics drops March 2026 job openings data on Tuesday, May 5. If job openings fall meaningfully below the 6.9 million February figure, it will signal that the low-hire equilibrium is deepening — and that talent is continuing to pool on the sidelines. That's a pipeline argument, not a pause argument.
The Bottom Line
Corporate America is in the middle of a mass restructuring premised on AI returns that, by the honest account of most executives leading the charge, haven't materialized yet. That doesn't mean AI won't deliver — it almost certainly will, and probably sooner than the productivity statistics currently suggest. But the companies making headcount decisions as if the returns are already in are taking a real risk: they're depleting the human infrastructure that turns AI investment into compounding advantage.
The Solow Paradox eventually resolved. Computers did transform productivity — it just took until the late 1990s for the gains to show up cleanly in the data. The same resolution is coming for AI. The recruiters and hiring managers who treat this period as a window rather than a wall will be better positioned when it does.
BlueLine's matching engine connects hiring managers with qualified candidates across industries and roles. If you're hiring through the current market, start here.