Developer Partners

Posted on May 29 • Originally published at developerpartners.com

Are Tech Layoffs Really Because of AI?

#ai #leadership #career

From the boardroom to the PR office, Silicon Valley has produced a favorite line to explain mass layoffs: "AI is changing the way we work." Companies including Google, Salesforce, and Amazon all used the expansion of AI to justify tens of thousands of job losses. The story is clean, forward-looking, and pleasing to investors. The question is: was it true?

When we examine the timelines and data, a different picture emerges, one worth exploring in good faith rather than cynicism.

The Data Tells a Different Story

One fact that rarely surfaced in discussions about AI-driven layoffs is that the wave of tech layoffs in the US began in early 2022, more than six months before ChatGPT became publicly available in November 2022. US tech companies recorded an estimated 93,000 layoffs in 2022, and that number surpassed 200,000 in 2023.

At the time those layoffs began, large language models such as ChatGPT were not yet available for enterprise use. The technologies that companies would later cite as the reason for restructuring their workforces were still prototypes or purely experimental, confined to academic papers and research labs. The timing alone should give us pause before accepting "AI productivity" as the driving explanation.

Meta offers a telling example. On November 9, 2022, three weeks before ChatGPT's public launch, Mark Zuckerberg announced the elimination of roughly 11,000 jobs, approximately 13% of Meta's workforce. The company cited macroeconomic headwinds, rising costs, and a slowdown in its advertising business. Then, in March 2023, after declaring that year Meta's "year of efficiency," Zuckerberg cut another 10,000 employees. Only at that point did AI begin to feature prominently in the company's stated rationale, despite there being little measurable AI-driven productivity gain at Meta to justify the decision.

This is not a minor discrepancy in timing. It is a fundamental problem with the narrative. You cannot credibly attribute job cuts to a technology that did not yet exist in deployable form at the time those decisions were made.

Where Is The AI Productivity Increase?

If AI had genuinely been replacing workers and driving productivity gains in 2023 and 2024, we would expect to see a corresponding increase in measurable output. The evidence does not support that conclusion.

Acemoglu (2024, "The Simple Macroeconomics of AI," NBER) found that AI's near-term productivity impact was likely modest, with only a small share of tasks being cost-effectively automated in the short run. A Goldman Sachs report (Hatzius et al., 2023) acknowledged long-run transformative potential while noting that significant productivity effects would take years to materialize at scale.

Workplace studies did show generative AI improving performance in writing, customer support, software development, and translation. Brynjolfsson, Li, and Raymond (2023, "Generative AI at Work," NBER) found meaningful gains for individual workers in customer-facing roles. But the gap between controlled test environments and real-world organizational complexity remained large. Enterprise-wide gains sufficient to justify broad workforce reductions had not been demonstrated.

The question for 2022-2024 is not whether AI will eventually deliver those gains, but whether it has already done enough to warrant the restructuring attributed to it. Corporate decisions made in 2022 and 2023 cannot be retroactively justified by productivity gains that researchers say had not yet materialized.

The Broader Economic Context Companies Preferred Not to Lead With

Looking at the economic environment between 2022 and 2024, the pressures driving cost-cutting decisions were substantial and largely unrelated to AI. Central banks raised interest rates aggressively to combat inflation. Energy prices surged. Russia's invasion of Ukraine in February 2022 disrupted global supply chains and created lasting geopolitical uncertainty. Consumer confidence fell. And venture capital funding dropped sharply from its 2021 peak: the same peak that had fueled an extraordinary hiring boom in tech during 2020 and 2021.

Now, in 2025 and 2026, a new and severe shock has compounded these pressures. On February 28, 2026, the United States and Israel launched military strikes against Iran under Operation Epic Fury, triggering the closure of the Strait of Hormuz through which approximately 20% of the world's seaborne oil normally flows. Gas prices surged above $4 per gallon nationally, rising more than 38% from pre-war levels according to AAA. Brent crude climbed above $100 per barrel. April 2026 inflation reached 3.8% annually: the highest since mid-2023, driven primarily by energy costs. Economists at Moody's Analytics and Oxford Economics have warned the damage will persist for months even after the conflict ends.

These conditions represent classic macroeconomic pressures that have historically led companies to reduce their largest variable cost: labor, which typically accounts for 65-70% of total operating expenses.

Many tech companies had nearly doubled their employee headcounts between 2019 and 2022, riding a pandemic-era surge in demand. The layoffs that followed were, in large part, a correction of aggressive over-hiring, something many executives eventually acknowledged directly. Amazon's CEO Andy Jassy said as much in his 2023 layoff memo, describing a company that had "hired rapidly" and needed to course-correct.

The Stock Market Motivation

Why would companies frame economically motivated layoffs as AI-driven restructuring? The financial incentives are well documented. Research by economists studying market reactions to layoff announcements (Farber, 2005; Hallock, 1998) has shown that layoffs attributed to efficiency improvements are rewarded by markets, while layoffs attributed to falling demand are penalized. In a 2023 analysis, one economist summarized the dynamic plainly: "A layoff announcement has traditionally been interpreted by the stock market as an indicator of weak demand for a company's product. However, announcing a layoff resulting from improved efficiency tends to be interpreted positively." This asymmetry creates a powerful incentive to use the AI narrative.

The data from specific announcements confirms the pattern: Meta shares climbed nearly 8% when it announced 11,000 job cuts in November 2022; rose approximately 1% on its second round of 10,000 cuts in March 2023; and soared roughly 81% across the full year as investors rewarded the cost-reduction narrative. Amazon's stock rose roughly 9% in the month following its announcement of 18,000 layoffs, and Google's increased approximately 15% after Alphabet announced 12,000 cuts. In each case, markets interpreted the announcements as signals of improved future profitability rather than evidence of underlying business weakness.

Capital Is Moving, Not Vanishing

Perhaps the most revealing aspect of the entire story is what companies did with the money saved by cutting workers. They did not return it to shareholders or sit on it. They redirected it toward AI infrastructure.

Companies were not laying off workers because AI had made those workers redundant. They were laying off workers in order to fund the AI investments they hoped would eventually make them competitive. The savings from labor cost reductions were being converted into capital expenditures on data centers, chips, and computer infrastructure.

The scale of this reallocation is extraordinary. Combined capital expenditures on AI infrastructure across Alphabet, Amazon, Meta, Microsoft, and Oracle rose from approximately $162 billion in 2022 to $448 billion in 2025. The four largest hyperscalers were on track to spend over $700 billion building data centers and AI infrastructure by 2026: a 70% year-over-year increase. Meta's trajectory illustrates this most starkly. The same company that cut 21,000 employees across 2022 and 2023 committed to spending between $125 billion and $145 billion on AI infrastructure in 2026 alone, nearly double its $72 billion spend in 2025. Its CFO confirmed during the Q1 2026 earnings call that a "leaner operating model" was explicitly designed to offset surging AI capital expenditures.

The layoffs were not a consequence of AI making people unnecessary. They were a funding mechanism for AI infrastructure that had not yet delivered the productivity gains being used to justify the cuts.

What Is Actually Driving Workforce Changes Today

Understanding 2025 and 2026 requires examining forces that go beyond the early years of this story.

AI capabilities have grown meaningfully since 2022, and that growth is now creating genuine uncertainty of a different kind. Companies are less sure what their workforces will look like in two years than at any point in recent memory. This uncertainty alone is reshaping hiring decisions. When the cost of a wrong hire is compounded by the possibility that a particular role may be automated within 18 months, companies default to caution. The result is slower hiring, narrower headcounts, and a pronounced shift toward contract and project-based work rather than permanent employment. For workers, this translates into a job market that feels unstable even when headline unemployment figures appear relatively contained.

This uncertainty is also fueling an investment dynamic that closely mirrors previous technological manias. Just as the dot-com boom drove everyone to buy internet stocks regardless of underlying fundamentals, and just as rising real estate prices in 2005 through 2007 incentivized overleveraged buying, the perceived transformative potential of AI is pushing companies to invest heavily before the competitive picture clears. Fear of falling behind is as powerful a driver as evidence of productivity gains. The risk, as with previous manias, is that investment levels outpace what the technology can actually deliver in the near term.

Macroeconomic pressure has not eased. Interest rates remain elevated, the Ukraine conflict continues to constrain European energy markets and global supply chains, and the Iran war's energy shock has pushed inflation back toward levels not seen since 2023. These conditions keep labor under pressure independent of anything AI is doing.

A quieter but significant structural shift is also accelerating: offshoring. The COVID pandemic normalized remote work at scale, and companies quickly recognized that geography no longer constrained their hiring. What began as a domestic flexibility experiment became a global cost arbitrage opportunity. Roles in software development, customer support, finance, and operations that once required local presence began migrating to lower-cost labor markets across South and Southeast Asia, Eastern Europe, and Latin America. This trend, active since 2021, has continued and in some sectors intensified through 2025 and 2026, quietly removing jobs from high-cost markets that AI is often credited with displacing.

The Uncertainty Factor

Two distinct forms of uncertainty have made this entire period hard to navigate and easy to obscure. The first was economic: high interest rates, energy volatility, and geopolitical conflict created an environment where companies defaulted to cutting variable costs. The second was technological: AI was a genuine strategic inflection point, but its timeline remained undefined. No one, including the developers of the most sophisticated systems, could reliably predict which jobs would be automated, how quickly, or what competitive advantage would accrue to early movers.

Both pressures were real. The problem was that "AI is changing the way we work" became a shorthand for a far more complicated reality: macroeconomic pressure, over-hiring corrections, capital reallocation, offshoring, and stock market incentives.

Final Reflections

Most companies that reduced their workforces in this period were not acting in bad faith. They were making difficult decisions in a genuinely difficult environment.

But framing matters. When "AI is replacing workers" becomes the dominant narrative for largely macroeconomic and strategic decisions, it shapes how the public, policymakers, and workers understand both the technology and the labor market. It makes AI seem more disruptive in the present than the evidence supports, and it obscures the actual forces: interest rates, over-hiring corrections, geopolitical shocks, offshoring, and capital reallocation that drove the decisions. It also risks generating policy responses calibrated to the wrong problem entirely.

The wave of layoffs that began in early 2022 was real. The economic pressures behind them were real. The capital reallocation toward AI infrastructure was real. What was not established, and what the evidence does not support, is that AI had already done enough, by 2022 or even 2024, to be the primary cause of the workforce changes attributed to it. In 2025 and 2026, the picture is becoming genuinely more complicated. AI capabilities are advancing, offshoring is accelerating, macroeconomic shocks keep arriving, and companies are investing at a scale that reflects deep strategic conviction regardless of near-term returns. The honest answer for workers and policymakers alike is that multiple forces are reshaping employment simultaneously, and no single narrative, including the AI one, is sufficient to explain all of them.

References

DEV Community