AI Job Cuts Are Rising

Oracle cut about 30,000 jobs this year. That's the single largest layoff announced in 2026, and it landed inside a much bigger wave. By late July, tech layoff trackers had counted roughly 168,000 job losses since January, spread across close to 470 separate events, averaging around 835 a day. That's not actually the worst year on record. 2025 cut more people in raw numbers. What's changed is the pace, and what companies are willing to say out loud about the reason.

The AI share is climbing, and fast

Challenger, Gray and Christmas tracks the stated reason behind every announced layoff in the US. AI showed up as a cause in just 0.6 percent of job cuts in 2024. That rose to 4.5 percent in 2025, then to 13 percent in the first quarter of 2026. By spring, the firm had counted close to 50,000 job cuts tied to AI in 2026 alone, out of roughly 300,000 total layoffs announced so far this year.

Some of the individual cuts are specific enough to name. Amazon eliminated 14,000 corporate roles, saying AI lets it run with a leaner structure. Workday cut about 1,750 jobs, roughly 8.5 percent of its staff, to shift spending toward AI. Salesforce brought its support team down from 9,000 people to 5,000. Klarna went the other way. It cut 700 support agents for a chatbot in 2023 and 2024, then the CEO admitted the company had gone too far, and started rehiring people.

The bigger story is who isn't getting hired

Look past the layoff headlines and a quieter shift is showing up in the hiring data, and it's aimed almost entirely at people early in their careers. A Stanford Digital Economy Lab study, using ADP payroll records, now finds employment for workers aged 22 to 25 in the most AI exposed jobs shrinking by 3.8 percent a year, up from a 2.8 percent decline in 2024. Across all ages, the same exposed occupations only contracted 0.2 percent. The damage isn't spread out. It's concentrated at the entry point.

A separate Harvard working paper, built from resume and job posting data covering 66 million workers across more than 280,000 US firms, found entry-level hiring down roughly 80 percent per quarter at companies that adopted generative AI, since 2023. Researchers behind these findings point out this isn't mass firing. It's a collapse in new openings. Young workers simply can't find the jobs that used to exist. No one gets a severance package for a job that was never posted.

Most people who lose their job don't blame AI

Here's the part that complicates the story. Gallup found that only 1 percent of laid off workers named AI or automation as the main reason for their job loss. Most people point to restructuring, or a role simply going away.

But Gallup also found that 62 percent of laid off workers were people who didn't use AI themselves. Tech workers who used AI once a month or less were three times more likely to have been laid off than workers who used it more often, 18 percent against 6 percent. Nobody is losing their job because a chatbot took it. They may be losing it because they never learned to work alongside one.

The people who built this are arguing about what it means

Dario Amodei, Anthropic's CEO, warned last year that up to half of entry-level white collar jobs could disappear within five years, and unemployment could climb to somewhere between 10 and 20 percent. In May, Sam Altman went the other direction. Speaking in Sydney, he said he no longer expects the "jobs apocalypse" he once warned about, and called himself delighted to be wrong.

The Yale Budget Lab backs part of that. It found no meaningful change in unemployment through March 2026 for workers in the jobs most exposed to AI. Boston Consulting Group still projects that up to 15 percent of US jobs could be eliminated over the next five years. MIT's Daron Acemoglu keeps pushing back on Stanford's Erik Brynjolfsson over how large the eventual productivity gains really are. Two economists, looking at similar data, landing in different places.

None of this adds up to a jobless world, not yet. What it adds up to is a narrower door at the start of a career, and a labor market that punishes people for standing still with a tool everyone else is already using. That's a smaller crisis than the headlines suggest. It may also be a harder one to fix, because nobody gets fired for a door that never opened.

This newsletter can't do that work for you. What it can do is keep putting the real numbers in front of you as they shift, and flag what's worth acting on before it becomes the next headline. The actual work, learning the tools, updating what you offer, having the harder conversations at your own company, happens in your hours, not in this inbox. Do that steadily for a few months and it tends to add up in ways one email never could.

Install, import your projects, and it runs quietly on macOS or Windows. Timeglass drafts accurate timesheets from your real work while you do the work. Start free.

Try the AI tracker