I build AI agents. Like, thats literally what I do all day — I wire up autonomous systems that scout the internet, write content, publish articles, and report back to me without me touching anything. So when I see headlines screaming about 90,000 tech workers getting fired because of AI, I have a very specific reaction, which is: some of these companies are telling the truth, and a lot of them are completely full of it.
The numbers are real though. Challenger, Gray & Christmas released a report this week showing 52,050 tech layoffs in Q1 2026 alone — thats a 40% jump from the same period last year. In March alone, AI was cited as the reason for 15,341 of those firings, which is 25% of all tech job cuts that month. A month earlier that number was 10%. So the trend is accelerating fast. TrueUp's tracker puts the running total even higher at around 90,000 tech workers impacted across 212 companies since January. And then Oracle dropped a bomb this week — somewhere between 20,000 and 30,000 employees got a 6 AM email telling them they were done. Their corporate Slack went from 165,000 users to 155,000 in a single day. The freed-up cash is reportedly going straight to AI data center investments where Oracle has a $20 billion funding gap.
But heres where it gets interesting. Marc Andreessen went on the 20VC podcast this week and basically called the whole thing "AI washing." His argument is that every large company overhired during the pandemic by at least 25%, some by as much as 75%, and now they're using AI as a convenient excuse to do the layoffs they should have done two years ago. And honestly? I think he's at least partially right. Salesforce CEO Marc Benioff said something similar — that companies are cutting workers for several different reasons and lumping them all together under the AI label because it makes the company look like theyre on the cutting edge rather than just cleaning up bad hiring decisions.
Think about it from a CEO's perspective for a second. If you fire 10,000 people and say "we overhired during COVID and our margins are terrible," your stock tanks and everyone calls you an idiot. But if you fire 10,000 people and say "AI is replacing these roles as part of our strategic pivot to artificial intelligence," suddenly youre a visionary. The narrative completely changes. Wall Street loves it. The board loves it. Your remaining employees are scared enough to work harder. Its the perfect corporate magic trick.
But I'm not going to pretend AI isnt actually replacing people too because it def is. Block laid off 4,000 people — 40% of their entire workforce — and Jack Dorsey was unusually honest about it. He said flat out that "this is not driven by financial difficulty, but by the growing capability of AI tools to perform a wider range of tasks." Amazon cut 16,000 corporate roles this year and heavily implied AI would handle the work. Meta's planning to cut up to 15,000 employees to offset their massive AI investments. Atlassian cut 10% of its workforce specifically because AI changes the skills they need across the business.
A Duke University survey of 750 CFOs published last month tried to quantify the actual impact and came up with a number that's simultaneously scary and underwhelming: AI could eliminate about 500,000 jobs from the US economy in 2026. Not 500,000 people getting fired, but 500,000 fewer jobs existing than would have existed without AI — through a mix of actual layoffs and companies just not hiring for roles they would have filled. Thats about 42,000 jobs a month, and the researcher who ran the study pointed out thats actually huge when the US is only adding about 10,000 jobs per month right now. So the net effect is real even if individual headlines are exaggerated.
The part that nobody wants to talk about is who's getting cut and who's getting hired. Anthropic, OpenAI, and xAI are hiring aggresively. Claude paid subscriptions more than doubled in 2026. OpenAI hit $25 billion in annualized revenue. The companies building AI are growing fast. The companies buying AI to replace their workforce are the ones doing the firing. And the jobs getting cut arent actually the senior engineers everyone worries about — the CFO survey found that companies plan to increase skilled technical roles like engineers and data scientists while decreasing routine clerical work like data entry. An Anthropic analysis found that programming, customer service, and data entry are the most exposed categories.
So whats actually happening is a rebalancing, not an apocalypse. The total number of tech jobs is shrinking, yes. But the jobs that are growing pay more and require different skills than the jobs disappearing. Thats cold comfort if you just got a 6 AM email from "Oracle Leadership" (not even a real persons name, just the company signing it like a robot, which is almost poetic), but its the reality of what the data shows. A Stanford Digital Economy Lab note put it well — there wont be a single moment where everyone realizes AI is eliminating jobs. It'll keep creeping up, and at some point we'll realize in hindsight that it already happened.
My actual take as someone building these systems: the companies that are being honest about AI replacing roles (Block, Amazon) are prob the ones you want to watch because they're actually integrating AI deeply enough that it changes headcount. The companies using AI as an excuse for bad management (you know who you are) will rehire in 18 months when they realize they cut the wrong people. And for developers — the demand for people who can actually build, deploy, and maintain AI systems has never been higher. The irony of the whole situation is that AI is simultaneously the thing killing jobs and the thing creating the most new ones. The question is just whether youre building the robots or getting replaced by them.
