Ten days into September, the tech industry had already eliminated more than 6,300 jobs — and crossed a line it did not cross at all last year. The tracker Layoffs.fyi counts 128,536 technology workers cut at 299 companies globally in 2026 through September 10, more than the 122,606 cuts across 278 companies recorded in all of 2025, with nearly four months still on the calendar.
The September wave was led by Uber, which said on September 2 it would cut about 3,300 corporate roles — roughly 10 percent of its workforce, and its largest reduction since the 6,700 jobs eliminated in May 2020. PayPal, Apple, Oracle and Zomato also cut in the same stretch, per a Business Standard tally published September 11.
The obvious explanation is that AI is taking the jobs. The evidence is thinner than the headline numbers suggest — and the companies doing the cutting are, in several cases, not making that claim themselves.
What the companies actually said
Uber's memo is instructive precisely because it does not blame AI. CEO Dara Khosrowshahi told staff that years of growth had produced "more layers, more coordination, more fragmented ownership, and in some cases structures that made sense when businesses were smaller but no longer serve us well at our current scale." The restructuring, he wrote, was "designed to do two things: make Uber simpler and faster, and create more capacity to invest in our future."
The mechanics bear that out. Uber cut employees sitting seven or more layers below the CEO by 20 percent and reduced "micro-teams" — those with one or two direct reports — by nearly half. This is a span-of-control exercise with a robotaxi capital plan behind it, not an automation story. Uber grew at double digits through the period.
PayPal's 220 India cuts, about 4 percent of its India headcount, are part of a turnaround targeting $400 million in savings by year-end and $1.5 billion over two to three years. Apple's roughly 200 August cuts — about 100 in Vision Pro, 100 across Siri and AI software, per Bloomberg — were a reallocation: Siri is being rebuilt around a new AI system requiring different expertise.
Oracle is the outlier. Its 21,000-job reduction in June, about 13 percent of its workforce, comes closest to a direct AI link, and arrived alongside a sharp increase in cloud and AI infrastructure spending. Layoffs.fyi ranks Oracle first for the year, then Amazon (17,267), Dell (11,000), Meta (10,400), Microsoft (4,800) and PayPal (4,760).
The causation problem
Challenger, Gray & Christmas, which tracks stated reasons for US job-cut announcements, has AI cited in 116,175 announcements year-to-date — about 22 percent of all cuts, and the leading single reason for the year. But in August, AI fell to fourth with just 3,462 cuts, its lowest monthly total since December 2025, ending a five-month run at the top. Restructuring led instead, with 16,173. Technology announced 6,103 cuts in August, its quietest month of 2026, though its year-to-date total of 155,126 is up 52 percent year-over-year.
"What we'd like to see with low layoffs is an increase in hiring activity," said Andy Challenger, the firm's chief revenue officer.
Labor economists are blunter about the attribution gap. A July 2026 policy brief from the Stanford Institute for Economic Policy Research, co-authored by SIEPR director Neale Mahoney and former Bureau of Labor Statistics commissioner Erika McEntarfer, found that unemployment among the most AI-exposed quintile of workers has risen 0.77 percentage points since 2022 — slightly less than the 0.85-point rise among the least exposed. "These aggregate trends suggest a broadly softening labor market, rather than one characterized by AI-driven job losses," the authors wrote. The Census Bureau puts firm AI adoption near 20 percent; only 5 percent of firms report any employment impact, split evenly between gains and losses.
Forrester vice president J.P. Gownder put the test plainly: "If you are laying people off without a mature, ready-to-go AI agent to do the work, you are not laying off people because of AI. You are laying people off for financial reasons and then imagining that at some future date, AI may be able to do the work."
Wharton's Peter Cappelli points at investors, not models. "The companies that are laying off are not struggling," he said. Cutting headcount "makes revenue per employee jump up quickly."
Chetan Mangalwedhe, CEO of hiring platform TalentiFi-X, offers a defensible middle estimate: "roughly a quarter to a third of this wave represents genuine task replacement through AI and automation. The rest is largely cost reset and workforce restructuring."
Where the damage is real
The clearest AI signal is not in layoffs but in hiring that never happens. New-graduate unemployment hit 5.6 percent in early 2026, up 1.6 points in three years. Stanford's 2026 AI Index shows employment among software developers aged 22 to 25 down nearly 20 percent since 2024, and Gartner found 22 percent of CHROs reported at least one leader halting entry-level hiring because of AI. When Erik Brynjolfsson and co-authors added controls for the 2022 rate-hike cycle and remote work, entry-level declines did not become notable until 2024 — when the models actually got good.
Meanwhile, capital is moving. Alphabet, Amazon, Microsoft and Meta are expected to spend roughly $725 billion on capex in 2026, up 77 percent from about $410 billion in 2025, per Instahyre co-founder Sarbojit Mallick. As Revelio Labs managing director Evan Sohn framed it: "Payroll is being converted into capital expenditure."
What to watch
Whether Challenger's AI attribution keeps falling — a sustained drop would suggest the label was always partly rhetorical. The entry-level pipeline: campus hiring, junior developer postings, whether the 22–25 cohort stabilizes. And reversals: Klarna and Duolingo both walked back aggressive automation after service quality suffered. If firms that cut on anticipated capability start quietly rehiring, that is the cleanest evidence yet that the AI in these announcements was a forecast, not a fact. For the 128,536 people already out, the distinction is academic. For the next cohort, it is everything.
“If you are laying people off without a mature, ready-to-go AI agent to do the work, you are not laying off people because of AI. You are laying people off for financial reasons and then imagining that at some future date, AI may be able to do the work.”— J.P. Gownder, Vice President, Forrester