May's spike puts 2026 US tech layoffs 16.8% ahead of last year's pace

The 16.8% rise amounts to 13,560 more layoffs than in the first eight months of 2025 ^[1]. May supplied 31,513 on its own, more than twice that gap and about a third of the year's count ^[2]. January, at more than 20,000 ^[2], and May together account for more than half of the eight-month total ^[10]. Two announcements supplied much of it: Amazon's 16,000-worker reduction in January ^[6] and Meta's 8,000 in May, 10% of its workforce ^[7], together 24,000, roughly a quarter of everything Crunchbase counted ^[3].

The summer decline is small by the same measure. A 16.2% drop to 19,331 implies about 23,070 layoffs in June through August of 2025, so this summer came in roughly 3,700 lower ^[4]. That is less than half of Meta's single May cut ^[5]. Crunchbase itself says it is too early to establish a lasting reversal ^[5]. Uber reportedly laid off 3,300 people, 10% of its workforce, in early September ^[9], about 950 more than the tracker recorded for all of August ^[6].

Roger Lee, who founded Layoffs.fyi, explains the bursts as budget reallocation: established companies spending heavily on AI and cutting costs elsewhere, hoping smaller workforces will be more productive ^[13]. Andrew Challenger, of Challenger, Gray & Christmas, described the same move inside a single company. "They're letting people go from one area of their organization while they might even be hiring in an area that is focused on AI," he said ^[16]. He also sees plain substitution. "There are jobs that are literally being replaced by artificial intelligence," he told Crunchbase News ^[15]. Lee, speaking of this year's largest AI-attributed cuts, disagrees. "There's been little evidence that AI is actually replacing the work of the human employees let go," he said ^[12].

The AI label has spread anyway. Layoffs.fyi found AI cited in 33% of tech layoff events this year, against 1% in 2024 ^[10], and attributes 92,913 layoffs worldwide, 72% of its total, to AI ^[11]. That implies a global count near 129,000 ^[8]. It also leaves Lee's AI-attributed figure only about 1,100 short of Crunchbase's all-cause US count ^[9].

Measurement is a separate problem. Crunchbase excluded a reported fall of about 21,000 in Oracle's workforce during its fiscal year ended May 31, 2026, because the count and timing of the cuts were unclear ^[8]. That figure alone exceeds the tracker's June-to-August total ^[7].

I think the bursts are real and belong to a small number of employers, or rather to the dates on which a few of them decide to announce. Lee said big companies made up "about 87% of everyone laid off in 2026," close to last year's 85% ^[14]. With that concentration, the summer figures describe three months in which no large employer announced a cut on Meta's scale. The trackers count cuts and not hires, and neither publishes AI spending, so they cannot show whether budgets are moving to AI or whether tech headcount is falling. The counter-case is that the reallocation was front-loaded into January and May and has mostly run its course, leaving the rest of 2026 below 2025. A steadier monthly count spread across mid-sized employers, with the big-company share falling well below 85%, would prove the concentration view wrong.