Data for the first quarter of 2026 is in: the tech industry laid off approximately 78,000 people in total, with 47.9% attributed to AI and automation. 76% of the affected positions were in the United States. Oracle alone cut more than 10,000 jobs, while simultaneously investing the savings into data center construction.
It looks like the story of AI replacing humans is unfolding.
But as soon as you start asking detailed questions, the story becomes much less clean.
How Much of the "AI Did It" Claim Is True?
The Chief AI Officer at Cognizant made a remark worth noting:
"Sometimes, you know, AI becomes the scapegoat from a financial perspective, like when a company hired too many."
In other words: sometimes, AI is just a financial scapegoat. Companies hired too many people in the last cycle and now need to contract. Using AI as a reason sounds better, more forward-looking, and harder to challenge than saying "we misjudged demand."
OpenAI's CEO has also publicly acknowledged this phenomenon: some companies use AI as a reason for layoffs they would have made anyway.
Harvard Business Review published an analysis in January with a title that gets straight to the point: tech companies are laying off workers more because of expectations of what AI "could do in the future" than because of what AI "is doing now." CEOs at Ford, Amazon, Salesforce, and JPMorgan have all publicly stated that many white-collar jobs will "soon" be replaced by AI. Companies are acting on that expectation in advance, rather than waiting until AI actually performs those jobs.
But Real Displacement Is Also Happening
It would be unfair to attribute everything to "scapegoating."
Some layoffs do directly correspond to the deployment of AI tools:
- Automated code review tools go live, reducing the workload for junior developers
- AI customer service systems are deployed, shrinking human support teams
- Content generation and report automation reduce the need for back-office functions
The data supports this as well: of the 78,000 positions cut in Q1, roughly 20% (about 9,000 jobs) can be clearly attributed to reduced labor demand from AI productivity gains. The rest are more a case of "going with the flow"—AI provides a narrative, but the reasons are multiple.
Distinguishing between "AI actually did the work" and "the company believes AI will do the work later" is key to understanding this round of layoffs.
The Other Side: The More Companies Invest in AI, the More They Hire
At the same time, AI-related job postings increased by 92%, with an average salary premium of 56%.
IBM did not lay off staff; instead, it tripled its entry-level hiring. The reasoning: AI needs people to supervise, maintain, and iterate, and these tasks require more human labor.
European data is even more interesting: companies that invest heavily in AI are more likely to expand hiring than their peers, rather than lay off staff—because they are building new capabilities, not simply replacing existing processes with AI.
This reveals two fundamentally different logics for companies regarding AI:
| Substitution Logic | Expansion Logic |
|---|---|
| AI does the work of humans; lay off the humans doing that work | AI enables companies to do things they couldn't do before, requiring more people to operate at a larger scale |
| Short-term cost reduction | Medium- to long-term capability building |
| Typical: Oracle, Dell restructuring | Typical: IBM, heavy AI-investing European firms |
Both logics are operating simultaneously, just distributed across different companies and different types of roles.
A 6 to 12 Month Time Lag
Researchers broadly agree that there is a 6 to 12 month lag between AI deployment and actual productivity gains.
Today's layoffs and the real value delivered by AI are misaligned on the timeline. Companies are cutting people now, but efficiency improvements won't show up in the numbers for six months to a year. If expectations are not met, these layoffs will leave real capability gaps that will need to be refilled by rehiring.
Of the 78,000 positions cut in Q1—how many were because AI actually did the work, how many because the company believed AI would do the work later, and how many were normal organizational adjustments made under the banner of AI—no one can give a clean answer.
That Cognizant AI leader's word "scapegoat" is really accurate.
The question is not whether AI will take jobs, but which types of jobs, at what point in time, and whether they are being replaced in reality or by expectation—only by distinguishing these dimensions can we be said to be seriously discussing the issue.
Sources: Tech industry lays off nearly 80,000 employees in the first quarter of 2026 — almost 50 percent of affected positions cut due to AI (Tom's Hardware); CocoLoop; Companies Are Laying Off Workers Because of AI's Potential—Not Its Performance (Harvard Business Review)