AI was supposed to kill engineering jobs, but new data suggests they’re the most resilient

Whether or not AI is already changing jobs is the topic of fierce debate.

Tech layoffs hit their highest single month whole in years in Might, and AI was the most-cited purpose, in accordance with outplacement agency Challenger, Grey & Christmas.

Software program engineering, in idea, is the skilled area most susceptible to automation, given the speedy adoption of AI-powered coding instruments. Nevertheless, researchers at enterprise agency SignalFire say the hiring information tells a unique story.

“The rationale given for plenty of layoffs is constantly AI, and particularly they’ll say AI with respect to code; they’ll say one engineer might do the job of nevertheless many engineers up to now,” stated Asher Bantock, SignalFire’s head of analysis. “What we’re seeing on the bottom is a bit inconsistent with that.”

SignalFire’s evaluation, which tracked the careers of thousands and thousands of staff throughout greater than 80 million corporations, means that engineering was essentially the most resilient job operate in 2025. As a substitute of specializing in layoffs, that are tough to trace as a result of individuals typically delay updating their employment standing after job cuts, SignalFire examined hiring information as a extra correct indicator of real-time workforce developments.

Whereas whole hiring throughout massive tech corporations dropped 25% in comparison with 2019 ranges, engineering roles noticed a a lot smaller decline of simply 11%, in accordance with SignalFire’s newest “State of Expertise Report.”

In truth, engineers comprised 55% of all new hires in 2025 throughout the 12 corporations SignalFire classifies as “Tech Majors” — Alphabet, Meta, Apple, Amazon, Microsoft, Netflix, Nvidia, Tesla, Uber, Airbnb, Block, and Stripe. This can be a important leap from 2019, when engineers represented solely 46% of recent recruits, in accordance with the report.

The continued want for engineers was much more evident at early-stage startups, which collectively introduced on 7% extra engineers in 2025 than they did in 2019, SignalFire’s information exhibits.

If AI had been actually substituting for engineering expertise, Bantock argued, engineering hiring could be the primary to fall amid the present tech hiring contraction. As a substitute, SignalFire’s information exhibits that engineering headcount is rising sooner than most different job features in tech.

Whereas Anthropic CEO Dario Amodei warned final 12 months that AI might wipe out half of all entry-level white-collar jobs and push unemployment as excessive as 20% inside 5 years, the corporate’s personal head of economics, Peter McCrory, advised Trendster in March that he had not but seen any important AI-driven results on the workforce.

Stated McCrory on the time: “There’s at the least no bigger materials distinction in unemployment charges” between staff who use Claude for the “most central process of their job in automated methods” — like technical writers, information entry clerks, and software program engineers — and staff in jobs much less uncovered to AI that require “bodily interplay and dexterity with the actual world.” 

Nvidia CEO Jensen Huang went additional nonetheless, outright rejecting the idea that AI will substitute engineers. “Any individual stated that AI goes to destroy all the software program engineering jobs,” Huang stated in an interview on the Stanford Graduate College of Enterprise in April. He then argued the other is true. Now that every one engineers at Nvidia are utilizing agentic AI, “software program engineers are busier than ever,” he’d stated.

Huang added that whereas brokers are writing code close to instantaneously, they’re continually pushing engineers to generate “the following thought.”

For now at the least, it appears that evidently armed with AI, engineering has turn out to be a basic instance of the Jevons paradox — the concept larger effectivity doesn’t cut back demand for a useful resource; it will increase it, as a result of the work expands to fill the brand new capability. As Bantock stated of engineering expertise on this second: “They’re instantly much more productive, and there’s infinite work for them to do.”