What's actually driving tech layoffs in 2026 (and what engineers should do about it)

Google reported $62 billion in net profit in Q1 2026. At the same time, software engineers across the industry were dealing with layoffs, slower hiring, and one of the most confusing job markets they’ve seen in years.

These two facts sit next to each other in nearly every conversation I have about the tech job market right now, and they create a genuine puzzle. 

Companies printing money at historic rates are cutting engineers and pointing at AI as the reason. The responses have split in two directions: one side says this proves AI is finally replacing engineers at scale, the other says the AI explanation is a cover story for something more mundane. Both sides have found evidence. Neither is quite right.

What's actually driving tech layoffs in 2026 is more specific than either narrative, and understanding it changes what engineers should do about it.

The financial mechanics behind the announcements

When a company generating tens of billions in quarterly profit announces it's cutting engineers, the obvious question is why? The answer these companies usually supply is that they're redirecting budget toward AI. That framing has been repeated often enough that it's become the default explanation for the current wave of tech layoffs.

It's not wrong, exactly. But it's doing less explanatory work than it appears to.

Capital expenditure on AI infrastructure is growing faster than most companies projected even a year ago. The major tech companies are significantly raising their AI spending targets, and that kind of announcement creates its own stock price problem. Spending more on something without a clear near-term revenue path is, by definition, a signal investors price cautiously. Markets don't love uncertainty.

Layoff announcements function, in part, as an offset to that uncertainty. The message to investors becomes roughly, we're spending more on AI, and we're reducing headcount costs to balance it. It presents a kind of financial net neutrality that cushions what would otherwise be a harder sell to the market.

This isn't a conspiracy. It's a signaling mechanism, and it's one that's been developing for a few years. Before 2023, large tech companies didn't typically announce layoffs publicly. They happened quietly, surfacing through leaks. Then companies started announcing them, partly because markets responded positively. Cutting people while profitable read as efficiency. Stock prices went up.

That playbook has grown more complicated. The most recent rounds haven't produced the same stock price lifts because investors are now weighing the layoffs against the AI spending they're meant to justify, and the math is murkier.

It's also worth noting that the raw numbers, in many cases, aren't as unusual as the headlines suggest. 

Companies like Meta have long maintained a practice of turning over roughly five to ten percent of their workforce annually through performance reviews, internal calibration, and the ongoing process of deciding who's in the right role. The current layoff numbers at many of the loudest companies aren't dramatically outside that historical range. What's changed is how those numbers are being framed and why.

That framing matters because it changes how engineers interpret their own risk. If layoffs are simply “AI replacing engineers,” the takeaway is fear. If they’re only investor theater, the takeaway is dismissal. The harder truth is that both the market story and the work itself are changing at the same time.

You can see that in how engineers are responding.

Three ways engineers are responding to AI

Among the engineers I work with and talk to, a pattern has become hard to ignore. Most engineers fall into one of three buckets, and which bucket you're in matters more right now than it has in a long time. 

The leading edge

A smaller group, maybe ten to twenty percent, is pushing AI tools as hard as they go. They're running multiple agents in parallel, writing less code by hand, and occasionally shipping things that break in new ways. They're sometimes ahead of what the tools can reliably deliver. But they're also building a working model of where this is heading, and that model is compounding.

The middle majority

A larger group, probably the majority, is keeping pace without pushing any envelopes. They're using the tools their company has made available, staying current with what they're told to adopt, and doing solid work. They're not inventing anything new about the workflow; they're following it.

The skeptics, and why this is the riskiest position

Then there's a third group that's actively skeptical. The argument runs something like: AI makes engineers feel more productive without actually making them more productive. The output looks right but it's brittle. I don't use AI for coding because it loses something the manual process has.

This group is carrying the most professional risk right now.

If AI development stopped today, if the models were maxed out and never improved further, the operational improvements already available would still be significant. Engineers who've built real fluency with these tools are already operating at a measurably different output level than those who haven't. The engineering in the leading-edge bucket will iron out their flows, and the majority middle will eventually catch up, drastically increasing their AI usage from today. The floor has moved, regardless of what happens next with the technology.

The debate over whether AI will eventually replace most engineering work remains open. The debate about whether refusing to use available tools is a sustainable professional strategy is less so. A calculator doesn't do the thinking. But an engineer who insists on doing the arithmetic by hand isn't demonstrating depth. They're creating friction.

The engineers in that third bucket aren't necessarily less skilled. In some cases they're reacting to something real. AI tools do produce plausible-looking output that's wrong in non-obvious ways, and the engineers who catch those errors are the ones who understand the underlying systems. That instinct is correct. The error is concluding that the right response is avoidance rather than calibrated use.

The point isn’t that every engineer needs to chase the most extreme AI workflow or turn themselves into a different kind of worker overnight. The better question is what AI changes about leverage. In a market where companies are trying to get more output from smaller teams, the engineers who stand out are the ones whose value is specific, visible, and hard to flatten into a generic headcount number.

That’s where the career question gets more personal.

If you're currently in a role: find your superpower

The advice that holds across all this uncertainty is structural, not tactical.

Think of an engineering team the way you'd think of a professional sports roster. A pitcher and a catcher are both baseball players. Ask either of them to cover another position and they'll probably do fine. But the pitcher is there because of one specific thing they do exceptionally well. That's leverage.

The engineers navigating this market most effectively are the ones who've identified their version of that position and consistently put themselves in situations that reinforce it rather than dilute it.

What a real superpower looks like today

That superpower almost never lives in a stack or a framework anymore. It tends to live somewhere more durable:

• When the system is down and everything is on fire, I'm the person the team wants in the room.
• I'm the one who can translate what we're building to people who aren't technical, in a way that actually helps the project move.
• I'm a raw output person, the one you call when a backlog needs to get done.

These capabilities travel. They compound. And they're the hardest to automate first, because they require judgment about what the output is even for. When you continue to invest in your superpowers, you become the person the AI companies want to train on, rather than the ones whose jobs are replaced.

Push back on pressure to be well-rounded

If you're in a role and your manager is pushing you to be more well-rounded rather than deeper in what you're already good at, that's worth pushing back on. The more valuable move, for you and for your team, is building around specific strengths rather than averaging them out. Talk to your manager about it. If the team structure doesn't support that, it's worth finding one that does.

For engineers earlier in their careers, finding that position is still partly about discovery. Try enough things, and keep the question active: could this be what I'm best at? For engineers further along, the harder task is often just naming a superpower that's already there. It's easy to describe yourself in the terms your last job title gave you. A mentor or a structured program can help translate that into something more precise.

If you've been laid off: precision beats versatility

The instinct under immediate pressure is to expand your target. Any company, any stack, any role, because more surface area feels like more opportunity. In a competitive market, that instinct typically backfires.

Companies hiring right now are hiring specifically. They want someone who fits a particular need, and a generic pitch is hard to land in any particular slot. Being clear about who you are and equally clear about who you aren't tends to lead to fewer initial conversations and better eventual fits.

Reframe around your actual strengths

We worked with an engineer at Formation whose last role involved significant work on a well-known open-source project: presenting about it at conferences, working with companies building integrations on top of it, representing the project's direction to external stakeholders. He was describing himself as a front-end engineer. Once he reframed around what he was actually doing, the ability to operate in large-scale open-source systems, to communicate technical decisions externally, to represent a project in public, the number of interviews he was getting per week changed significantly. The skill set was identical. The positioning was different.

A role you've contorted yourself into to move quickly might be a start. It's rarely the stable version of the next thing. The goal is a job where your actual strengths are what's being evaluated, and getting specific about what those are, even when the pressure to be broadly available is real, tends to produce better outcomes faster.

Junior hiring isn't coming back quickly

Junior hiring is likely to stay slow. The category of bounded, trainable work that once justified bringing in someone new is shrinking as senior engineers, using AI tools, absorb more of it. This is a structural change, and the industry hasn't worked out what replaces the old apprenticeship model at scale. The senior engineers of 2030 will have to come from somewhere. Where exactly is an open question.

Where roles are expanding

There are areas of genuine growth. Forward-deployed engineers, people who work directly with clients to help them use and integrate increasingly complex software, are seeing growing demand partly because the software being produced is outpacing the ability of most organizations to absorb it. Security-adjacent roles are seeing renewed investment. The definition of what an engineering team looks like is still being renegotiated at most companies.

The role itself is changing

The job of writing code as a standalone professional identity is almost certainly changing. What replaces it as the center of the engineering role isn't settled. Engineers like Andrej Karpathy have described workflows built around running multiple AI agents in parallel and intervening mainly at the level of architecture and direction. That's not yet ordinary everywhere. But it's no longer strange enough to dismiss.

The engineers I see positioned most calmly for what's ahead aren't the ones who've predicted the outcome. They're the ones who've gotten specific about what they're best at, and who are preparing in ways that match how companies actually evaluate candidates today. That's not a formula for certainty. It's a posture for navigating genuine uncertainty without waiting for it to resolve.

Sign up for our newsletter

Get the latest in tech right in your inbox

Prepare for the market as it actually is

If you're working through what your position is, or preparing to make a strong case for it in a competitive hiring process, Formation's Fellowship pairs you with mentors who understand how the market is actually moving.