In the earnings presentations of the big tech companies, there is a pattern that is repeated with such regularity that it almost no longer generates headlines. The company announces that it is going to invest billions in artificial intelligence. He then announces that he is laying off thousands of employees. He then explains that both things are related: the money saved on payroll goes to investment in AI. And finally, because investors demand it and press releases demand it, add something about how AI will create new job categories in the future.
That promised future is still future. What is present, countable and verifiable are the 92,000 technology sector employees who have lost their jobs in the first five months of 2026, according to the records of the specialized portal Layoffs.fyi. The number does not include adjacent sectors — media, banking, legal services — where AI-driven streamlining is also producing significant cuts. The number is just the tech. And it's just so far this year.
Microsoft and Meta: 16,000 layoffs in a single day
The most symptomatic episode of the pattern was when Microsoft and Meta announced layoffs on the same day, adding between the two more than 16,000 affected employees. Meta laid off 8,000 workers, 10% of its global workforce, and left another 6,000 vacancies that were in the selection process unfilled. Microsoft, for its part, launched a mechanism that the company had not used in its 51 years of history: incentivized voluntary dismissal. Instead of announcing collective cuts, it offered attractive exit packages so that it was the employees themselves who decided to leave.
The business logic of both decisions is understandable on their own terms. The two companies are investing tens of billions in AI infrastructure — data centers, chips, model development — and need to balance that investment by reducing operating costs in other areas. Support employees, administration, some less specialized software engineering profiles and middle management roles are the most vulnerable to this restructuring logic.
Tech layoffs in 2026 — selection of cases (until May)
- Total global technology sector: more than 92,000 employees (Layoffs.fyi)
- Goal: 8,000 employees (10% of staff) + 6,000 canceled vacancies
- Microsoft: incentivized voluntary layoffs — first time in 51 years
- Oracle: 30,000 layoffs, 2.1 billion in reserved severance payments
- DeepL: 25% of staff eliminated to "integrate AI at all levels"
- Total Big Tech investment in AI during 2026: estimated at $700 billion
DeepL and the extreme case
If there is a case that starkly illustrates the logic of the moment, it is DeepL. The German company that created the most used translation tool in the professional world announced that it would cut a quarter of all its employees. The stated motive, in the words of its founder, was to "profoundly transform the way the company operates by integrating AI at all levels of its organization."
DeepL is, literally, an artificial intelligence company. Its core product has always been a neural translation engine. You are not "adopting" AI from the outside; AI is what it has always been. What the announcement reveals is that even existing AI companies are using more advanced models to replace tasks previously done by people within their own organizations. It's AI eating AI.
The real cost of layoffs
There is one element of the official narrative of tech layoffs that rarely makes headlines: layoffs are extraordinarily expensive. Oracle set aside $2.1 billion for severance alone in its round of 30,000 layoffs. Microsoft has committed similar payments for its voluntary departures. Taken together, estimates suggest that large technology companies are spending tens of billions on severance pay, legal costs of dismissal processes, loss of productivity during transitions, and the cost of hiring new profiles that replace them.
In purely short-term financial terms, many analysts point out that mass layoffs rarely produce the savings they promise. What they produce is a signal: a signal to investors that the company is serious about the transformation towards AI, that it is willing to make difficult decisions, that management has strategic clarity. It is financial communication as much as operational restructuring.
AI today allows small groups to perform tasks that previously required entire teams. What no one discusses with enough honesty is what happens to the teams that previously performed these tasks.
When does the promised job arrive?
The official narrative of big tech companies and AI advocates is that each wave of automation destroys jobs in the short term and creates more, generally better-paying jobs in the long term. The industrial revolution, the mechanization of agriculture, the industrial automation of the 1980s and 1990s: in all of these cases, the economy ended up creating more work than it destroyed, although the transition was painful for those who lived through it.
The problem with applying that argument to the current moment is that the speed of the transition is qualitatively different from that of previous ones. When the mechanization of agriculture displaced workers from the countryside, it took decades and generated a migration process to cities and factories that, although difficult, was navigable. When generative AI displaces a service, analytics, communications, or programming worker, the time available to adapt is measured in months, not decades.
The World Economic Forum identifies in its latest projections that the demand for new roles related to AI is real and growing. It also identifies that these roles require skills that most displaced workers do not have and cannot acquire in the time they have available. The gap between "AI will create jobs" and "those jobs will be available to people who lost theirs" is the question that no tech company has satisfactorily answered.
Share this article
