← All articles

§ ARTICLE / Deep Dive

The Layoff Story We're Telling vs. the One That's True

AI, Overhiring, and the Correction We Mistook for Disruption

When Block cut 40% of its workforce and cited AI, its stock surged 24%. That single data point explains more about the current layoff wave than any automation thesis. Tech companies have a powerful financial incentive to frame workforce reductions as strategic AI pivots rather than admissions of pandemic-era overhiring — and the market has rewarded them for it. But the data tells a different story: the roles being eliminated look far more like a cyclical correction than a structural shift. Less than 1% of recent layoffs are genuinely caused by AI automation. Mistaking this correction for the real disruption isn’t just inaccurate — it leaves organizations unprepared for when AI-driven restructuring actually arrives.

The Convenient Narrative: Why AI Gets Blamed

Block announced a 40% workforce reduction in early 2025 — more than 4,000 people gone — and explicitly cited rapidly improving artificial intelligence as the driver [9]. Its stock surged 24%. Meta’s stock climbed on reports of a planned 20% cut tied to AI spending. The pattern repeated across the sector: invoke AI, get rewarded by investors. There’s a straightforward reason for this. A workforce reduction framed around AI adoption sends a signal to investors that a straightforward cost-cutting announcement does not [1]. It transforms a layoff from an admission of poor planning into a story about strategic foresight. The financial incentives are enormous. AI-related stocks have accounted for roughly 75% of S&P 500 returns since ChatGPT’s launch [1]. Seventy-nine percent of U.S. CEOs fear losing their jobs within two years if they don’t deliver measurable, AI-driven business gains [1]. When your compensation is tied to stock performance and your board wants an AI narrative, framing layoffs as AI-driven isn’t dishonesty — it’s rational corporate behavior. Sam Altman himself acknowledged the dynamic, noting that some companies are blaming AI for layoffs that would have happened regardless [10]. CNBC reported on the phenomenon under the label “AI-washing” — companies attributing job cuts to artificial intelligence when the real drivers are cost reduction, restructuring, or simple overcapacity [7]. The market initially bought it. But there are signs the skepticism cycle has begun. Goldman Sachs noted a shift in 2025: investors started interpreting AI-driven layoffs as a negative signal, with affected companies experiencing an average 2% stock price drop [1]. The AI-washing premium is eroding. The narrative, however, has already done its work. Headlines conflate correlation with causation. Layoff announcements in tech continued to mount in March 2026, up 24% year-over-year, and coverage routinely describes AI as the catalyst [6]. The frame is sticky, even as the evidence underneath it frays.

The Pandemic Hiring Surge by the Numbers

Between 2019 and their peak headcounts, the largest tech companies expanded their workforces at rates that dwarfed anything in their histories. Meta nearly doubled its headcount. Alphabet added tens of thousands of employees. Amazon’s workforce exploded — driven partly by logistics, but also by a hiring philosophy that assumed the pandemic’s demand patterns were permanent [2][5]. This wasn’t limited to the tech giants. Across the sector, companies raced to capitalize on anticipated shifts to e-commerce and remote work [2]. The hiring spree was fueled by near-zero interest rates and a growth-at-all-costs capital environment that punished caution and rewarded expansion. Headcount growth outpaced revenue growth. Roles were created to pursue initiatives that made sense in a world where every human interaction was mediated through a screen — and made less sense as that world receded. The correction should have arrived in 2023. Instead, it was delayed by a combination of optimism, institutional inertia, and the first wave of AI excitement that made companies reluctant to cut headcount while they were still figuring out what large language models might mean for their businesses. By the time the cuts came, the gap between headcount and sustainable workload had widened further. The research compiled for our analysis at Deerfield Green puts this plainly: tech layoffs attributed to AI are largely a correction of COVID-era overhiring, with AI efficiency serving as narrative cover for restructuring that was inevitable regardless of AI advances [2]. The magnitude of the excess hiring created a correction that was always going to be painful. The only question was what story companies would tell about it.

What the Layoff Data Actually Shows

If AI were driving the current wave of layoffs, you’d expect the cuts to concentrate in roles most susceptible to automation — data entry, routine analysis, content generation, customer service. That’s not what the data shows. Instead, the positions eliminated skew heavily toward the roles that expanded most aggressively during the pandemic: recruiting, middle management, project-based expansion roles, and coordination functions that existed primarily to manage the growth itself [2]. These are not the roles where large language models are replacing human judgment. They’re the roles that exist when a company is growing fast and disappear when it isn’t. Our research found that less than 1% of recent layoffs are genuinely caused by AI automation [3]. The overwhelming majority are corrections to pandemic-era overhiring, repackaged under an AI efficiency narrative that rewards companies with higher stock prices [3]. This distinction matters enormously. A recruiting coordinator laid off because the company has stopped growing isn’t the same as a copywriter replaced by GPT-4. One reflects a cyclical business decision. The other reflects a structural change in how work gets done. Conflating them produces bad strategy. Entry-level hiring is down 25-50% or more in some datasets [1], which could reflect either dynamic — but the timing aligns far better with the end of the expansion cycle than with the maturation of AI tools. The AI systems that could genuinely displace entry-level knowledge work at scale are still being deployed, still being measured, still being integrated into workflows. The layoffs, meanwhile, have already happened. The gap between narrative and evidence is where the real strategic risk lives. Companies that believe they’ve already restructured for AI may think they’re ahead. They’ve mostly just corrected their headcount to pre-pandemic baselines.

Where AI Is Genuinely Reshaping Work

None of this means AI isn’t affecting employment. It means the mechanism is different from what the headlines suggest. Genuine AI displacement is beginning at the margins — particularly in customer service and content translation [2]. But the more pervasive effect isn’t layoffs. It’s what doesn’t happen: the position that doesn’t get backfilled after someone leaves. The team that stays the same size while handling 30% more volume. The hiring plan that gets scaled back not because AI can do the job today, but because a manager believes it might be able to next quarter. This is the distinction between AI causing a layoff and AI changing how you think about the next hire. The first is a discrete event that shows up in layoff data. The second is a slow-moving pressure that shows up in headcount growth rates — or rather, the absence of them. A company that would have added five customer support agents this year adds two instead, because the existing team augmented with AI tools can handle the projected volume. Nobody was fired. Nobody will appear in a layoff tracker. But three jobs that would have existed now don’t. This is where AI’s real impact on employment is accumulating. It’s harder to measure, harder to headline, and harder to plan for — which is exactly why it deserves more attention than the dramatic but misleading layoff narrative. The productivity gains are real but uneven. Some teams see 40-60% improvements in specific workflows. Others see modest efficiency gains that don’t yet justify restructuring. The variance is high, and most organizations are still figuring out where the gains are durable versus where they’re illusory. The companies making strategic decisions based on the assumption that AI has already restructured the labor market are building on faulty premises. The ones tracking backfill rates, hiring velocity, and job composition changes are closer to the truth.

Strategic Implications: Preparing for the Real Disruption

If the current layoffs are mostly cyclical correction — and the evidence strongly suggests they are — then the real AI-driven restructuring hasn’t happened yet. Organizations that treat this moment as the main event will cut too deep, too fast, and then pay to rebuild capacity they shouldn’t have shed. The Deerfield Green framework for distinguishing cyclical from structural layoffs rests on three diagnostic questions. First: are the eliminated roles concentrated in growth-supporting functions like recruiting and program management, or in core operational functions where AI tools are demonstrably replacing human output? If it’s the former, you’re looking at a correction. Second: is the company reducing headcount while maintaining or increasing output in the affected areas, or is it simply doing less? If output stays flat or grows, structural change may be real. If it contracts, you’re watching a business shrink, not transform. Third: is the company investing in the AI infrastructure, training, and workflow redesign that genuine automation requires — or just cutting costs and calling it strategy? Real AI integration requires capital and deliberate effort. Layoffs without corresponding investment are just layoffs. For workforce planners, the actionable insight is this: build your strategy around the slow pressure of reduced backfill and altered job composition, not around the dramatic but misleading layoff narrative. Invest in skills that complement AI tools rather than competing with them. Track the leading indicators — hiring velocity, job posting composition, internal mobility patterns — rather than the lagging ones. And recognize that the companies currently announcing AI-driven layoffs are, in many cases, the same ones that overhired during the pandemic and are now using a compelling story to excuse a correction they should have made two years earlier. The real disruption is coming. It will look different from this one.

The current layoff wave is a story about poor pandemic-era planning, not about AI replacing human workers. That distinction isn’t academic — it has direct consequences for how organizations allocate capital, design workforce strategies, and prepare for what’s next. Companies that mistake a cyclical correction for a structural shift will over-index on cost cutting and under-invest in the integration work that genuine AI transformation requires. They’ll emerge leaner but not stronger, having shed capacity they’ll need to rebuild at premium rates. The AI-driven restructuring of work is real, but it’s arriving slowly, unevenly, and through mechanisms — reduced backfill, altered job composition, slower headcount growth — that don’t generate headlines. The organizations that will navigate it well are the ones watching the right signals: not layoff announcements, but hiring patterns. Not CEO statements about AI strategy, but actual investment in tools, training, and workflow redesign. The correction we’re in now was inevitable. The disruption still to come will be optional only for those who see it clearly enough to prepare.

References

Deerfield Green