American technology companies cut more than 38,000 jobs in May, the most in a single month in nearly two years, and the people doing the hiring say one factor dominates the decisions: artificial intelligence. The cuts bring the year-to-date total to nearly 124,000, up 66 percent from the same period last year, according to data compiled by a firm that tracks corporate layoffs.
The May figure, 38,242, was the highest monthly total since August 2024 and marks an acceleration of a trend that has quietly reshaped the industry. What is new is the stated reason. Employers that once described AI as a tool to help workers now describe it as a replacement for them, with companies citing AI in an increasing share of layoff notices as they automate programming, customer service and content production.
The change is visible in the job categories being cut. Software developers, who were largely protected in earlier waves of tech downsizing, now appear in restructuring plans alongside support staff. Companies building AI systems are the most aggressive: several firms that announced layoffs this spring told employees that the work being eliminated would be handled by models and agents rather than rehired, according to people who attended the meetings.
The aggregate picture is broader than technology. U.S. employers overall announced more than 97,000 layoffs in May, the highest May total since 2020, when the pandemic shut down the economy. That suggests the AI-driven cuts are not contained to the tech sector; banks, media companies, retailers and business services firms are all experimenting with models that can draft documents, answer calls and generate content.
The data fits a pattern labor economists have watched for two years. Early in the AI boom, the technology was framed as a complement to human work, a co-pilot that made employees faster. The current wave of layoffs reflects a harder calculus: for a class of routine knowledge work, the marginal cost of a model is approaching zero, and companies are beginning to treat headcount accordingly.
Executives are candid about the reasoning in private. At a technology conference this spring, several chief executives said they had paused hiring for roles that AI could plausibly perform within a year, and some said they were working through which teams could be reduced as the models improved. The decisions are economic, they argue, not ideological: if a customer-service interaction costs cents to automate and dollars to staff, the outcome is predetermined.
Workers are adapting unevenly. Demand remains strong for engineers who build and deploy AI systems, for data scientists and for salespeople who can sell the new tools, and companies report difficulty filling those roles even as they cut elsewhere. The result is a bifurcated labor market within tech itself: rising pay at the top of the AI stack, and growing insecurity for the routine positions that models can now handle.
The shift has begun to draw attention in Washington. Lawmakers have held hearings on AI and employment, and some states have floated proposals for wage subsidies or retraining programs, though none has advanced far. The policy response is complicated by the fact that the disruption is arriving faster than any previous technological transition on record, leaving the institutions built to manage such changes unprepared.
Economists caution against overreading a single month. Layoff announcements do not always become actual job losses, and the labor market has repeatedly proven more resilient than the headlines suggest. Some of the May increase reflects ordinary corporate restructuring that happens every year, and the technology sector has a history of cyclical cutbacks that reverse when the cycle turns.
The cuts are concentrated in the functions where AI is most capable. Programming, customer support, content production and data processing appear again and again in the restructuring plans, according to the layoff notices and internal memos reviewed. The pattern is telling: these are exactly the jobs that involve converting instructions into outputs, the kind of work language models have learned to do. The correlation between AI capability and job cuts has become hard to ignore.
The transition is creating a two-speed workforce. Demand for engineers who can build and deploy AI systems remains intense, and companies report difficulty filling roles that require working directly with the new models. At the same time, the entry-level and routine jobs that once served as on-ramps into the industry are disappearing, which economists say could hollow out the middle of the labor market. The result may be an industry that hires fewer people but pays the ones it keeps considerably more.
Economists are debating whether this is a repeat of past technology transitions or something new. Earlier waves of automation replaced tasks but created new jobs nearby, and the labor market eventually grew. The difference now, some argue, is that AI can perform judgment work, not just routine work, and that the new jobs being created are fewer than the ones being removed. The debate will shape how governments respond, but for the workers laid off in May, the distinction is academic: the jobs are gone, and the data says AI is the reason.
But the composition of the cuts points to something structural. When companies cite AI as the reason for eliminating positions, they are signaling that the jobs will not return when the economy improves, because the work itself has been redesigned around machines. The 66 percent jump in year-to-date layoffs is a measure of that redesign. The labor market is being rebuilt around a smaller workforce of people who manage AI and a larger share of work that AI does directly, and the pace of that reconstruction accelerated sharply in May.
