AI is supposed to be killing engineering jobs. The hiring data says engineers are the most resilient role in tech.

Ask the internet whether software engineering jobs are going away and you’ll get a confident yes. Layoff trackers are full of cuts blamed on AI. Executives talk about one engineer now doing the work of ten. The May 2026 numbers from outplacement firm Challenger, Gray & Christmas showed tech layoffs hitting their highest single-month total in years, with AI as the most-cited reason.

Then there’s the data on who actually got hired. It points the other way.

SignalFire, a venture firm that tracks employment across more than 80 million companies, just published its State of Talent Report 2026. The headline finding: engineering was the most resilient job function in tech last year. Not the most endangered. The most resilient. So the question isn’t really whether software engineering jobs are going away. It’s why the hiring data and the layoff headlines tell such different stories, and what that gap means if you’re the one looking for the job.

What the 2026 tech hiring data actually shows

SignalFire’s research head, Asher Bantock, put the disconnect plainly to TechCrunch. “The rationale given for lots of layoffs is consistently AI, and specifically they’ll say AI with respect to code; they’ll say one engineer could do the job of however many engineers in the past. What we’re seeing on the ground is a little inconsistent with that.”

Here’s the inconsistency, in numbers. Total hiring across large tech companies dropped 25% compared with 2019 levels. Engineering roles dropped only 11%. So yes, tech hiring contracted. But engineering held up more than twice as well as the rest of the org chart.

The share data is more striking. Across the 12 companies SignalFire calls “Tech Majors” (Alphabet, Meta, Apple, Amazon, Microsoft, Netflix, Nvidia, Tesla, Uber, Airbnb, Block, and Stripe), engineers made up 55% of all new hires in 2025. In 2019, that figure was 46%. The pie got smaller, and engineers took a bigger slice of it.

Early-stage startups went further. They collectively brought on 7% more engineers in 2025 than in 2019. Not a smaller decline. An actual increase, measured against a pre-pandemic baseline.

Bantock’s logic is hard to argue with: if AI were genuinely replacing engineering talent, engineering hiring would be the first thing to fall during a contraction. It’s the opposite. Engineering headcount is growing faster than most other functions in tech.

Why the layoff story and the hiring story disagree

Both things can be true at once. Companies are cutting jobs and citing AI, and companies are hiring engineers at a higher rate than before. The reason they look contradictory is that layoffs and hiring measure different moments, and one of them is much easier to see.

Layoffs come with press releases, WARN notices, and a wave of LinkedIn posts. Hiring is quieter. SignalFire actually chose to study hiring data over layoff data for a practical reason: people often delay updating their employment status after a job cut, which makes layoff figures laggy and noisy. Hiring is the cleaner real-time signal. And the cleaner signal says demand for engineers is intact.

There’s also a gap between what executives predict and what the same companies report seeing. Anthropic CEO Dario Amodei warned in 2025 that AI could wipe out half of all entry-level white-collar jobs and push unemployment as high as 20% within five years. But Anthropic’s own head of economics, Peter McCrory, told TechCrunch in March 2026 that he hadn’t yet seen significant AI-driven effects on the workforce. His words: “There’s at least no larger material difference in unemployment rates” between workers who use Claude for the most central tasks of their jobs (technical writers, data entry clerks, software engineers) and workers in jobs that require physical interaction with the real world.

So the company whose CEO gave the scariest forecast is also the company whose economist says the effect hasn’t shown up in the numbers yet. That’s worth sitting with before you accept the layoff narrative at face value.

The Jevons paradox, and why busier doesn’t mean fewer

Nvidia CEO Jensen Huang rejected the replacement theory outright. “Somebody said that AI is going to destroy all of the software engineering jobs,” he said at the Stanford Graduate School of Business in April 2026, then argued the reverse. With every engineer at Nvidia now using agentic AI, he said, “software engineers are busier than ever.” The agents write code almost instantly, which only pushes engineers to come up with the next idea faster.

Economists have a name for this. It’s the Jevons paradox: when a resource becomes more efficient to use, demand for it tends to rise rather than fall, because the work expands to fill the new capacity. Cheaper, faster engineering doesn’t mean companies need less of it. It means the backlog of things worth building gets longer, because more of them are now affordable to build.

Bantock described the current moment the same way. Engineers are “suddenly a lot more productive, and there’s endless work for them to do.”

This reframes the AI software engineer jobs 2026 question entirely. The threat was supposed to be that AI does the work, so the worker isn’t needed. What the data suggests instead is that AI does part of the work, the cost of building drops, and the appetite for building grows to match. The person who can direct the tools is more valuable, not less.

None of this means the field is easy right now. Hiring is still down 11% from 2019, entry-level roles are genuinely harder to land, and “engineers are resilient as a category” is cold comfort if you’re three months into a search with nothing to show for it. The point is narrower than “everything is fine.” It’s that the demand is real, and it’s being obscured by a louder story about disappearance.

The same companies cutting jobs are hiring for different ones

The clearest sign that “are software engineering jobs going away” is the wrong question comes from the companies on the layoff lists themselves. They aren’t shrinking their engineering function. They’re swapping out one set of skills for another.

IBM is the sharpest example. TechCrunch’s running list of 2026 AI layoffs notes that IBM has cut somewhere between 3,000 and 9,000 U.S. positions this year, and replaced roughly 200 HR roles with AI agents. The same entry cites Bloomberg reporting that IBM plans to triple its U.S. entry-level hiring for AI and hybrid-cloud roles. Cuts in one column, a tripling in another, at the same company, in the same year.

The pattern repeats down the list. General Motors trimmed IT roles but still had about 80 open positions in AI, autonomous vehicles, and related work, per CNBC. GitLab cut 14% of staff specifically to fund AI infrastructure and rebuild its platform for agent-scale workloads. The headcount that’s leaving tends to sit in support, middle management, and routine operations. The headcount that’s growing sits in AI, machine learning, and infrastructure.

For a job seeker, that changes what to optimize for. The roles tied to building, deploying, and running AI systems are the ones companies are protecting and expanding, even while they announce cuts elsewhere. Engineers who can show they work fluently with these tools, rather than around them, line up with where the open seats actually are. “Resilient as a category” is the average. Within that average, the AI-adjacent roles are doing better than the average, and the routine ones worse.

The roles exist. They’re just not where you’re looking.

Here’s the practical problem with resilient-but-quiet demand: it doesn’t always turn into a job posting. A startup that grew its engineering team 7% didn’t necessarily run 7% more public listings. Teams hire through referrals, through people they already know, through a hiring manager reaching out to someone whose work they saw. The job board is the last and most crowded place that demand shows up, if it shows up there at all.

That’s the gap between the data and your experience of the search. The hiring data says engineers are wanted. Your application tracker says you’ve sent 80 resumes into the void. Both are accurate. The demand is real; the application funnel is just a terrible way to reach it. And it’s getting worse, because the same AI that’s making engineers productive is also letting every other applicant fire off hundreds of applications, while AI screening tools filter most of them out before a human ever looks.

It’s also worth being precise about what the layoff headlines actually represent. When a company cuts roles and grows its AI usage in the same quarter, that’s often a reshaping of the org rather than a pure shrink, and reshaped teams are still hiring, just for different work and through different channels.

The way to reach a quiet opening is to go straight to the person who owns it. Find the engineering manager building the team. Look at what they shipped recently, what they’ve said about where the team is headed, what problem they’re clearly trying to solve. Then send a short, specific message about how you fit that problem. That route skips the application pile entirely, and it works in exactly the situations the public listings miss: the team that’s hiring but hasn’t posted, or posted and got buried under 400 auto-submitted resumes.

What to do with this

The data is pretty clear: engineering demand survived the AI scare, and the roles are real even when the postings aren’t. The bottleneck isn’t whether companies want engineers. It’s whether you can reach the ones doing the hiring before your resume disappears into a queue. Direct outreach is how you do that. angld.AI automates the research-to-outreach pipeline: paste a job or name a company, and it finds the hiring manager, researches them, and drafts a personalized message in about a minute, so you spend your time on the conversations that move a search forward instead of refreshing an application portal.

The layoff headlines say software engineering jobs are going away. The 2026 hiring data says they’re the most resilient role in tech. The opportunities are out there. They’re just sitting on teams that aren’t always posting, behind hiring managers you can reach directly. Applying alone won’t surface them. Reaching out will.