AI Didn't Cause Most of These Tech Layoffs. Stop Blaming the Robot.
AI became the most convenient scapegoat in modern corporate history.
From a communications standpoint, it is perfect.
"We are eliminating 8,000 roles as we transition to an AI-first operating model" sounds inevitable. It sounds strategic. It sounds visionary. It sounds like something that would have happened no matter who was sitting in the executive seats.
"We hired 8,000 people we did not actually need between 2020 and 2022, and now the interest rate environment has changed, and investors want profitability" sounds like what it actually is: a business that overextended during a historic capital bubble and is now paying the bill.
One of these narratives is much easier to put in a press release.
It also conveniently removes accountability from the executives who made the hiring decisions in the first place.
Calling layoffs an “AI transformation” makes cost reduction sound visionary instead of reactive.
The Actual Numbers
In 2025, AI was cited as the official reason for approximately 55,000 tech layoffs, according to Challenger, Gray & Christmas.
That number gets repeated constantly.
What gets repeated much less often: 55% of companies conducting layoffs in 2025 cited economic uncertainty as the primary driver, according to Resume.org. And separate analyses found that many of the roles being eliminated, recruiters, project managers, customer support teams, and middle management layers, are not roles AI is currently capable of replacing in any meaningful way.
So two things are happening simultaneously.
Some AI displacement is absolutely real and measurable, particularly in junior development, QA, and certain support functions.
But most of the headline layoff numbers are the delayed correction from a hiring binge that ran from 2020 through 2022, when capital was cheap, and growth was the only metric that mattered.
Blaming AI for all of it avoids the much more uncomfortable question:
Why were all those people hired in the first place?
How the Hiring Binge Happened
During the pandemic years, tech companies hired at a pace that made very little long-term sense.
Interest rates were near zero. Capital was cheap. Venture funding was everywhere. Every investor wanted growth. Every board wanted expansion. Every executive team wanted to prove they were capturing market share faster than competitors.
Nobody wanted to explain to investors why their company was growing slower than everyone else during the easiest capital environment in more than a decade.
So companies expanded aggressively:
- more recruiters
- more managers
- more experimental product teams
- more parallel initiatives
- more hiring simply because competitors were hiring too
The binge lasted roughly two years.
Then the Federal Reserve started raising interest rates.
The cost of capital increased. Growth-at-any-cost stopped being a convincing investor story. Profitability suddenly mattered again. Operational efficiency became the mandate. And thousands of roles that only made sense inside a zero-interest-rate growth bubble no longer had a defensible business case attached to them.
That is not AI.
That is a market cycle behaving exactly the way market cycles always behave.
What AI Is Actually Doing to Employment
None of this means AI is not affecting employment.
It absolutely is.
The data on junior developer hiring is particularly stark. Employment for developers aged 22 to 25 fell nearly 20% from its 2022 peak. That is real displacement documented across multiple credible data sources.
But the honest picture is more layered than the press releases suggest.
AI is genuinely automating repetitive and highly structured tasks:
- boilerplate code
- basic QA scripting
- structured data entry
- repetitive documentation
- certain categories of content generation
Roles built primarily around those tasks are shrinking.
That part is real.
But in many organizations, AI is not replacing entire roles. It is compressing the amount of labor required for specific categories of work.
Those are related ideas.
They are not the same thing.
What AI is generally not replacing is:
- engineering judgment
- architecture decisions
- stakeholder management
- organizational knowledge
- client relationships
- operational leadership
- cross-functional coordination
In other words, most of the work happening in senior-level technical roles.
When a company lays off 200 senior engineers and cites AI productivity gains, that is almost never literally about AI replacing those engineers.
It is about headcount reduction being the fastest lever available when margins need to improve.
The AI explanation sounds cleaner.
It is also usually incomplete.
Why the Narrative Sticks Anyway
Companies benefit enormously from the AI framing.
It makes difficult business decisions sound like inevitable technological transitions instead of executive miscalculations.
"We made aggressive hiring decisions during a zero-interest-rate bubble and now need to reverse them" sounds reckless.
"We are restructuring around AI" sounds innovative.
Nobody made mistakes.
The AI just happened.
There is also a real phenomenon happening in parallel that makes the narrative believable: AI tools are genuinely increasing the productivity of experienced developers. One senior engineer with strong AI tooling can absolutely produce more output today than the same engineer could two years ago.
That changes headcount math.
It just does not explain most of the 127,000 tech workers who lost jobs in 2025.
Wall Street historically rewards cost reduction much faster than it rewards long-term organizational stability. Companies know this. Executives know this. Investors know this.
So layoffs framed as “AI transformation” often receive a much more favorable reaction than layoffs framed as “we overexpanded during a growth bubble.”
Employees also do not push back very hard on the AI framing because it feels plausible. AI anxiety already exists. Productivity gains are visible. Entry-level hiring really is collapsing in some areas. The industry already feels unstable.
So the explanation feels emotionally true even when it is only partially true.
And once the outcome is unemployment, most people stop arguing about the wording.
The story stays simple even when reality is not.
What This Actually Means
If you are a developer or tech professional trying to read the market right now, the useful frame is this:
Not everything labeled as AI disruption is actually AI disruption.
The structural risk from AI is real in specific areas:
- entry-level development
- QA
- support
- repetitive operational work
- highly standardized task-based roles
Those are worth paying attention to and preparing for.
The skills gap between people who use AI tools effectively and people who do not is widening and will probably continue widening for years.
But most senior engineers laid off over the last three years did not lose their jobs because a language model replaced them.
They lost their jobs because their companies overhired during a historic growth cycle and eventually had to correct.
That is a different problem with a different solution.
The solution to “companies overhire during boom cycles” is:
- building financial runway
- maintaining portable skills
- understanding that boom periods are temporary
- not mistaking two good years for permanent stability
The solution to “AI is actively compressing my role category” is understanding which parts of your work are vulnerable to automation and adapting toward the parts that are not.
Conflating those two realities produces bad career decisions.
The tech layoff narrative of 2025 has a villain.
And the villain is AI.
It is a convenient story.
It is not the whole story.
Most of what happened was boring old business: companies that grew too fast, operated as though cheap capital would last forever, and eventually had to reverse decisions that only made sense inside a temporary economic environment.
AI is changing the industry.
That part is real.
But a lot of executives discovered something else at the same time:
“AI transformation” sounds much better in an earnings call than “we massively overhired and now need to cut costs.”
One of those stories sounds inevitable.
The other sounds avoidable.
If this hit close to home, subscribe to the newsletter for more honest conversations about software development, leadership, AI, and the realities nobody likes putting in investor decks.
Comments ()