AI Did It. Honest.
Image by Easy-Peasy.AI, licensed under CC BY 4.0
There is a new tradition in Silicon Valley, and it goes like this. You wait until the quarterly numbers are looking a bit ropey, you call an all-hands, and you tell everyone the company is pivoting to become “AI-first” and that regrettably this means saying goodbye to a meaningful percentage of the workforce. You then go on CNBC and talk about the future.
It’s a good trick. Layoffs caused by missed targets sound bad. Layoffs caused by artificial intelligence sound like you’re doing something visionary.
Let’s have a look at who’s been at it recently.
GitLab: The Agentic Era (Their Words, Not Mine)
GitLab announced in May 2026 that it was cutting 7% of its workforce, flattening three layers of management, reducing its country footprint by 30%, and reorganising R&D into 60 “autonomous teams.” CEO Bill Staples announced that GitLab was preparing for the “agentic era.”
The agentic era. Brilliant. Nothing says “we’re not just cutting costs” like coining your own epoch.
Here’s the thing though. GitLab’s AI product, Duo, only reached general availability in January 2026. GitHub Copilot launched in 2021 and has 20 million users. Cursor hit $2 billion in annualised revenue by February 2026. GitLab didn’t arrive late to the AI party. They arrived after the party had been going for five years, discovered the drinks were almost gone, and decided to blame the guests for leaving early.
And in the same restructuring, GitLab quietly retired its entire CREDIT values framework and with it “Transparency” and “Diversity, Inclusion and Belonging.” They announced this on their own blog on the same day as the layoffs, explaining they were “not retiring them because they were wrong, we are choosing instead to focus on something different for this era which demands a different operating posture.” The new values are Speed with Quality, Ownership Mindset, and Customer Outcomes. The company that built its entire reputation on radical openness dropped transparency from its values at the precise moment it needed to be transparent about why it was laying people off. The Hacker News thread noticed immediately. You genuinely couldn’t write it.
The stock dropped 8% immediately after the announcement. The market, at least, wasn’t buying the agentic era story.
Atlassian: Same Playbook, Different Jira Ticket
Two months earlier, in March 2026, Atlassian cut 1,600 people (roughly 10% of its workforce) citing the need to adapt to the AI era. The announcement was so similar in structure and language to GitLab’s that you’d be forgiven for wondering if both companies hired the same PR firm.
To be fair to Atlassian, they make tools for managing work and AI genuinely is changing how teams manage work. But 1,600 people is a lot of people. If AI were really doing that much of the job already, you’d expect the product to show it. Whether it does is left as an exercise for anyone who’s sat through a Confluence migration.
Coinbase: 14% Gone, One Email, Cheers
Coinbase cut 700 jobs in May 2026, about 14% of its workforce. CEO Brian Armstrong explained that “the pace of what’s possible with a small, focused team has changed dramatically, and it’s accelerating every day.” He also mentioned, cheerfully, that non-technical teams are now shipping code to production with AI assistance.
This is true! It is also what you say when you want to sound like Elon Musk rather than someone who just had a difficult quarter in crypto markets.
Three days later, Coinbase went down. For two hours. Customers couldn’t trade. To be fair, the culprit was an AWS datacentre where the chillers failed and a server room overheated, so it wasn’t the non-technical teams and their AI-assisted code that did it. But the timing was so perfectly bad that Gergely Orosz, who writes The Pragmatic Engineer and has seen a thing or two, noted on X that it was “unfortunate optics” to have an hours-long outage days after your CEO had been publicly talking about non-technical people shipping to production. That’s putting it diplomatically.
To be fair, Coinbase at least didn’t claim to be entering a named era. Points for restraint.
Oracle: We Made Them Train Their Own Replacement
This is the one that made me put my tea down.
Oracle has reportedly been laying off staff while simultaneously asking those same employees to train the AI systems that would replace them. Former Oracle employees shared stories this April of being let go by a single email after decades of service, some facing visa complications, others losing unvested stock in the process.
Training your replacement before being shown the door is the sort of thing you’d expect in a dystopian novel, not in the HR policy of one of the largest enterprise software companies on earth. Though I suppose if you’re going to go out, at least you get to say you built the thing. Cold comfort, that. Very cold.
The Bit They’re Not Telling You
Right, here’s the stat that should be getting more attention. A Gartner study of 350 global business executives found that 80% of those who cut jobs citing AI saw no increase in return on investment as a result. None. They cut the people, the ROI didn’t improve, and they’re presumably now hoping nobody asks too many questions.
Helen Poitevin, VP Analyst at Gartner, was pretty direct about it: “Looking only at layoffs is shortsighted in terms of getting value from AI. That’s not where the value is. That’s not where the productivity gains are going to be.”
And Sam Altman, who you’d expect to be the last person to downplay AI’s impact on jobs given he runs OpenAI, acknowledged that some companies are engaging in what he called “AI washing” by attributing layoffs to AI when the real motivations are somewhere else entirely. He estimated that fewer than 1% of 2025 job losses could be directly attributed to AI.
Less than one percent. So while company after company invokes the robots as cover, the cuts are mostly about overhiring during the easy money years, rising interest rates, compressed valuations, and the correction that was always going to come after 2021.
Sometimes, Yes
Sometimes, yes. Salesforce genuinely reduced its customer support headcount from 9,000 to 5,000 using agentic AI. That’s real. AI agents handling tier-one support queries is not science fiction, it’s already happening. If your job was answering the same twelve questions on rotation, the threat is genuine and the companies saying so aren’t lying.
Cloudflare: An Honest One
Cloudflare is the other case worth taking seriously, and it’s the one that gives you an actual framework rather than a press release. On 8 May 2026, CEO Matthew Prince laid off more than 20% of the workforce — 1,100 people — while simultaneously posting record quarterly revenue of $639.8 million, up 34% year-on-year. He was unusually specific about who went and why. His word for it was measurers: middle managers, internal audit, finance, operations, compliance — the people whose job is to count things, report on them, and oversee. His argument was that AI can now do that measuring continuously, objectively, and at a scale no human team could match. What AI is not coming for, in his telling, are builders and sellers. “If an engineer on my team can now be 10 times as productive,” he said, “I’m going to hire as many as I can find.”
That’s a meaningfully different story from the others in this list. Prince wasn’t blaming AI for a difficult quarter. He was growing at 30% and still cut a fifth of the company because the internal measurement and oversight layer had been made redundant by tools that do it better. You can disagree with the decision, but you can’t accuse him of using AI as cover. He named the category, explained the logic, and said out loud what most executives in this situation won’t.
How to Tell the Difference
But look at where the cuts are actually landing. When companies cut management layers, country footprints, and back-office functions all at once and call it AI, that’s restructuring. AI is just better optics than saying “we hired too many people when money was cheap and now we have to fix it.”
The tells are there if you look. Did the company’s AI product actually ship before the layoffs? Are the roles being cut the ones AI can genuinely automate, or are they the ones that were always going to go in a downturn? Did the CEO coin a new era name in the press release? Did they, and I cannot stress this enough, remove Transparency from their company values?
AI is changing things, no question. But right now “AI did it” is being used to explain an awful lot of decisions that had nothing to do with AI and everything to do with a bill that’s been coming for a while. Some of these layoffs are real transformation. Some are cost cutting with better branding. And some involve asking people to train the thing that’s about to make them redundant, which, frankly, takes some nerve.