In recent years, layoffs at technology companies have stopped being extraordinary events.
They have become part of the calendar.
With every new announcement, the script seems familiar: a company releases its financial results, talks about efficiency, restructuring, strategic focus, and artificial intelligence. Shortly afterward, hundreds or thousands of professionals discover that their positions have been eliminated.
For those watching the market from the outside, the conclusion seems simple:
Artificial intelligence is replacing technology professionals.
For those working in the industry, another explanation emerges:
Companies hired too aggressively during the pandemic and are now correcting the excess.
Both statements contain some truth.
And both, when used in isolation, conceal a much larger transformation.
Today’s layoffs are not caused by a single factor. They are the result of excessive hiring, economic changes, pressure for profitability, internal restructuring, and a technology that promises to allow fewer people to produce more.
AI did not create all these problems.
But it arrived at the perfect moment to accelerate them and, in some cases, justify them.
The most convenient narrative
Every layoff needs an explanation.
A company is unlikely to say publicly:
We hired without proper planning, created unnecessary structures, and now need to correct our decisions.
The language is usually more polished.
The company is “simplifying operations.”
It is “reducing layers.”
It is “reallocating resources.”
It is “prioritizing strategic areas.”
Or it is “preparing for the age of artificial intelligence.”
These statements may reflect real changes. The problem is that they also function as corporate language capable of turning management mistakes into technological inevitability.
When a company attributes layoffs to AI, it creates the impression that the decision could not have been avoided.
Leadership did not make a mistake.
Planning did not fail.
The strategy did not change.
The future simply arrived.
In this way, artificial intelligence risks becoming a universal justification for decisions that have far more complex causes.
That does not mean its impact is fictional.
It only means that we should be cautious before accepting any corporate statement as a complete analysis of the market.
Before AI, there was the pandemic
To understand the layoffs, we need to go back a few years.
During the pandemic, technology companies experienced extraordinary growth in demand for digital services. Remote work, e-commerce, online entertainment, cloud services, and communication tools became even more central to the economy.
Many companies interpreted that emergency-driven growth as a permanent shift.
They hired rapidly.
Expanded products.
Created new teams.
Launched experimental projects.
Increased administrative structures and management layers.
Capital was readily available, interest rates were low, and expectations for growth seemed almost unlimited. In that environment, hiring too slowly appeared more dangerous than hiring too much.
Then the situation changed.
In-person life partially returned. Growth in some services slowed. Interest rates increased. Capital became more expensive. Investors began demanding profitability, not just expansion.
Projects that once seemed strategic became expendable.
Teams built to sustain rapid growth no longer matched the companies’ new pace.
The first major waves of layoffs in the sector began before generative AI had been meaningfully integrated into business operations.
Therefore, attributing the entire current crisis to artificial intelligence would be historically inaccurate.
Some of the cuts are indeed a correction of the excessive optimism of the pandemic.
But that explanation is also becoming incomplete.
A correction should eventually come to an end.
The cuts have not.
AI does not need to replace a person directly
When we imagine AI replacing jobs, we tend to picture a machine doing exactly the same work as a person.
In practice, the impact often happens in a less visible way.
A company does not need to replace ten developers with ten autonomous agents.
It can give AI tools to eight developers, raise their productivity targets, and decide that the other two positions are no longer necessary.
AI can also accelerate testing, documentation, data analysis, internal support, contract review, prototyping, and code production.
None of these changes needs to eliminate an entire profession.
It is enough to reduce the number of people considered necessary to handle a given volume of work.
That is why the question “Can AI already replace a developer?” may lead us to the wrong conclusion.
Perhaps it still cannot take full responsibility for a system.
Perhaps it does not understand every business rule.
Perhaps it produces errors, vulnerabilities, and technically fragile decisions.
Even so, it can change the financial calculation used to determine the size of a team.
AI does not need to be perfect.
It only needs to convince the company that fewer professionals can produce similar results.
The numbers show that AI is already part of the justification
In the United States, companies announced 443,604 job cuts between January and June 2026. The technology sector accounted for 139,156 of them, an 83% increase compared with the same period in 2025.
In the same report, artificial intelligence was cited as the reason for 101,743 cuts, approximately 23% of all announced layoffs in the first half of the year. In June alone, AI appeared as the justification for 31% of the cuts. (challengergray.com)
These figures require two simultaneous interpretations.
The first is clear: artificial intelligence is already influencing decisions about workforce size.
The second is equally important: most cuts are still not directly attributed to it.
Economic conditions, business closures, lost contracts, acquisitions, and restructuring continue to appear among the leading causes.
Reality cannot be reduced to a headline claiming that “AI is stealing jobs.”
But it can no longer be reduced to the comfortable argument that “this is all still just a correction of pandemic-era excess.”
The real impact may be in budget allocation
There is another way AI can lead to layoffs without directly automating a position.
It competes for budget.
Building artificial intelligence infrastructure is expensive. Training and operating models requires chips, data centers, energy, storage, networks, and specialized teams.
Companies are investing enormous amounts of money in this race. To sustain those investments and demonstrate returns to shareholders, they need to reduce costs elsewhere.
This creates a transfer of resources.
Money that once funded certain projects, products, or teams is redirected toward infrastructure, software acquisition, and new AI-related initiatives.
A company may still be growing.
It may remain profitable.
It may continue hiring in certain areas.
And it may still eliminate thousands of positions elsewhere.
That pattern has appeared in recent announcements. In July 2026, Visa said it would eliminate approximately 2,600 positions, equivalent to around 7% of its workforce, with a particularly significant impact on technology and product teams. The company linked the restructuring to efficiency efforts and investment in growth areas, including AI, while stating that the technology was not the sole cause of the cuts. (reuters.com)
Microsoft used similar language when it announced approximately 4,800 cuts in July. The company said the eliminated roles would not be directly replaced by AI, but acknowledged that the technology is changing how work is performed. (uol.com.br)
These statements are not necessarily contradictory.
A position does not need to be occupied by a machine to be eliminated because of a strategy built around machines.
Efficiency for whom?
The word most frequently repeated in layoff announcements is probably not “intelligence.”
It is “efficiency.”
Companies say they need to operate more efficiently, reduce bureaucracy, accelerate decision-making, and bring leadership closer to execution.
All of that sounds reasonable.
The problem is that efficiency can mean different things.
For a technical team, efficiency may mean automating repetitive tasks, improving processes, reducing incidents, and delivering better products.
For the company, it may mean increasing revenue per employee.
For financial markets, it may mean maintaining or improving results with a smaller payroll.
These interests are not always aligned.
When a tool increases productivity, there are several ways to use the gains.
The company can reduce working hours.
It can improve quality.
It can develop new products.
It can eliminate repetitive work.
It can train its employees.
Or it can simply raise targets and reduce headcount.
Technology does not determine which path will be chosen.
That is an economic and managerial decision.
This is why the debate about AI and employment cannot be limited to the technical capabilities of the models. We also need to discuss who controls the productivity gains and how those gains will be distributed.
When ten people begin producing the equivalent of fifteen people’s work, the question is not only what will happen to the other five.
We also need to ask what will happen to the ten who remain.
Will they have better working conditions?
Will they work fewer hours?
Will they earn more?
Or will they simply be pressured to produce like fifteen people for the same salary?
The elimination of middle layers
Another important trend is the reduction of administrative structures.
During periods of expansion, companies did not hire only developers. They also increased the number of managers, coordinators, project leads, operations professionals, and other intermediary roles.
As cheap capital disappeared and pressure for efficiency increased, these structures came under scrutiny.
AI reinforces this movement by making it easier to produce reports, monitor metrics, document decisions, perform initial planning, and communicate across departments.
That does not mean leadership has become unnecessary.
It means companies are trying to operate with fewer layers between those who make decisions and those who execute the work.
In some cases, this simplification may remove real bureaucracy.
In others, it merely transfers responsibilities to professionals who are already overloaded.
The developer now writes code, speaks with the client, organizes the project, documents decisions, monitors metrics, and supports colleagues.
The company removes a layer.
The work does not disappear.
It is redistributed.
The market is laying people off and hiring at the same time
One of the most confusing aspects of this moment is that layoffs coexist with new hiring.
This only appears contradictory when we treat “technology” as a single category.
Companies are not necessarily abandoning the sector. They are changing priorities.
They eliminate teams connected to products with lower growth potential and hire professionals in data, security, cloud computing, AI infrastructure, model integration, and automation.
They also reduce generalist roles while competing for professionals capable of connecting artificial intelligence to real business problems.
In the United States, software development job postings increased by nearly 15% between February 2025 and mid-2026, following a prolonged period of decline. However, 71% of that growth came from senior positions, while 37% was driven by roles that mentioned AI in the job title itself. (hiringlab.org)
So we are not simply witnessing the destruction of the technology job market.
We are seeing a change in the composition of demand.
Some profiles are gaining ground.
Others are losing bargaining power.
And professionals who once met market requirements are discovering that the entry standard has changed.
What about Brazil?
We need to be careful when applying the reality of large American companies directly to Brazil.
The Brazilian market has a different structure.
A large number of companies still face basic challenges involving digitization, systems integration, security, infrastructure, data, and process automation.
In many businesses, the challenge is not to replace a highly technological operation with AI.
It is to organize systems that still depend on spreadsheets, manual processes, and scattered information.
That creates opportunities for technology professionals.
A forecast published by Brasscom estimates the creation of approximately 33,000 formal jobs in Brazil’s broader ICT sector during 2026, mainly in software and services. At the same time, the data shows a slowdown in formal job creation and growth in employment through individual contractors and other forms of work. Between 2023 and 2025, more than half of the recorded expansion occurred outside Brazil’s traditional employment framework. (datacenterdynamics.com)
The Brazilian market, therefore, has not simply stopped hiring.
But it is hiring differently.
There is a stronger preference for experienced professionals, an increase in service-based work, pressure on salaries, and growing difficulty for those seeking their first opportunity.
There is also a difference between net job creation and the individual experience of someone looking for a position.
A sector may end the year with positive job growth and still experience thousands of layoffs, employee replacement, and deteriorating working conditions.
A positive number does not mean everyone is safe.
Just as a series of layoffs does not mean the entire sector is dying.
AI can also become a scapegoat
There is a risk that companies will use artificial intelligence as a justification before they achieve any real productivity gains from it.
The promise is powerful.
Presentations show cost reductions, productivity gains, and automated operations. Leaders feel pressure to prove that they are not falling behind.
So investments are approved.
Headcount reduction targets are established.
Teams are cut.
Only afterward does the company discover that integrating AI into real systems is far more difficult than producing an impressive demonstration.
Data quality is poor.
Processes are not documented.
Legacy systems do not communicate.
Models produce incorrect answers.
Operating costs increase.
Legal and security risks emerge.
Professionals need to review almost everything.
In this situation, a company may eliminate positions based on productivity gains that have not yet been proven.
Months later, it discovers that the work still exists.
The usual result is overload, loss of institutional knowledge, declining quality, and new hiring under different job titles.
AI is not necessarily a lie.
But expectations surrounding it may be far ahead of its practical capabilities.
Not every layoff represents professional obsolescence
This point needs to be repeated.
A person is not laid off only because they are no longer competent.
Excellent professionals can be affected by financial decisions, product shutdowns, changes in direction, mergers, or leadership mistakes.
The idea that “only those who fail to adapt lose their jobs” would be a cruel oversimplification.
Adapting increases the chances of remaining relevant.
It does not create immunity from corporate decisions.
There is a difference between preparing for the market and blaming professionals for every negative outcome.
We can defend the importance of continuous learning without turning every layoff into an individual failure.
We can recognize that AI changes professions without accepting that companies should use that transformation to hide poor decisions.
We can encourage productivity without concluding that every smaller team is automatically more efficient.
How to interpret the cuts
When a company announces a new round of layoffs, it is worth asking several questions.
Is the company facing declining revenue, or is it still growing?
Is it shutting down entire products, or cutting professionals across multiple areas?
Is it reducing investment, or redirecting resources toward AI?
Have the roles genuinely become unnecessary, or will the work be distributed among those who remain?
Has the company demonstrated real productivity gains, or is it acting on expectations?
Is it training professionals for new roles, or simply replacing people?
These questions help separate technological transformation from cost-cutting presented as innovation.
We will not always have every answer.
But asking the right questions already prevents us from accepting superficial explanations.
So, who is to blame?
Artificial intelligence?
The excesses of the pandemic?
Interest rates?
Investors?
Executives?
The most honest answer is: it depends on the company, the moment, and the role.
In some organizations, layoffs are still correcting excessive hiring.
In others, they represent the elimination of projects that failed.
Some companies are cutting staff to fund AI investments.
Some functions are genuinely being automated.
Some leaders are using AI to justify financial targets that already existed.
And some companies are trying to copy their competitors without understanding what they are doing.
Artificial intelligence is not the only cause.
But it is no longer just a spectator.
It acts as an automation tool, a promise of productivity, a destination for investment, and an argument for restructuring.
Perhaps the correct question is not whether layoffs are the fault of AI or the pandemic.
Perhaps we are witnessing something larger:
a complete reassessment of how much human work companies believe they need in order to produce technology.
For years, growth meant hiring.
Now, growth may mean increasing output without increasing team size.
In some cases, it may mean producing more with fewer people.
That is the real turning point.
The technology job market is not disappearing.
But the relationship between growth, hiring, and professional security is being rewritten.
The pandemic created the excess.
Economic change demanded the correction.
And artificial intelligence offered companies a new possibility:
not merely returning to their previous size, but trying to operate with even smaller structures.
Blaming AI alone would therefore be too easy.
Ignoring its influence would be naive.
Mass layoffs are the result of human decisions made within a new technological context.
And perhaps that is the most uncomfortable conclusion.
Machines are not running companies yet.
The people deciding who gets laid off are still human.