U.S. software developer job postings are up nearly 15% since February 2025, and senior roles account for most of the gain

The prediction was simple: AI writes the code, developers get cut. The posting data so far says something messier.

U.S. software development job postings have risen nearly 15% since late February 2025, according to Indeed Hiring Lab. Over the same stretch, overall job postings fell about 7%. February 2025 is also when Anthropic launched Claude Code, which makes the timing hard to ignore. It does not prove the tools created the openings.

A rebound from a deep hole

The recovery is not a return to the good old days. Software development postings are still about 27.5% below their pre-pandemic level, per Indeed.

They also climbed off a steep drop from the 2022 hiring peak, when tech companies pulled back after a pandemic-era hiring binge. McKinsey Global Institute research notes that the earlier decline was not solely an AI story. Rising postings and rising AI use happening together is not evidence that one produced the other.

Who is getting hired

Senior roles accounted for 71% of the increase in software development postings between May 2025 and May 2026, Indeed found. Job titles that mention AI accounted for 37% of it.

Those are shares of the increase, not shares of all openings. They also cover a different window than the 15% figure, which runs from February 2025.

Indeed economist Sneha Puri said companies still need developers to oversee automated coding, refine products and help clients fit technology into daily operations. Demand is strongest, she said, for people who can help businesses apply AI tools.

The people locked out

The young are not sharing in the rebound. Stanford University's 2026 AI Index found employment among software developers ages 22 to 25 fell 20%. Across AI-exposed fields, employment for that age group fell 16% since generative AI took off, while older developer roles expanded.

Those are employment counts, not postings, which is why a rise in advertised jobs can sit beside a squeeze on entry-level hiring.

The code keeps piling up

GitHub reported in October that one in three pull requests, the mechanism developers use to propose changes to a project, now involves an AI agent. A year earlier the figure was fewer than one in 10. Some AI experts claim AI-written code across the whole platform has passed 50%. GitHub has not confirmed that.

GitHub went further in its own forecast: "If that pace holds, within the next two years, most of the code pushed to GitHub could be written by an agent. Much of it may never be fully read by a human."

That last sentence is where the jobs argument lives. GitHub also warned that more code means more security exposures, and said developers must prevent more of them and cut the human effort needed to fix the rest, or the volume of vulnerabilities could become untenable.

GitHub's 2025 Octoverse report counted nearly one billion commits pushed in 2025, up 25.1% from the year before.

The optimistic reading

Eric Kutcher, McKinsey's North America chair, put the upbeat case on CNBC's "Squawk on the Street." "It wasn't that long ago that we were reading the headline that there will be no more need for software developers," he said. Companies are producing more code than ever, he said, and developers who work with AI remain in high demand. "We're seeing much more growth in the places we thought we were gonna see the biggest displacement."

The argument has a historical analogue. Writing in CIO.com, one technology commentator invokes the Jevons paradox: when a resource gets cheaper to use, total consumption can rise rather than fall. Cheaper code, on this view, makes previously unaffordable projects worth building. The same piece cites Faros AI's developer reports, which found AI speeds up individual coders but review and rework swallow the gains before they reach the team. The commentator adds that the Jevons paradox is not a promise that demand always rises.

What the numbers do not settle

Postings measure what employers advertise, not what they hire. They say nothing about pay, and a 15% rise off a depressed base leaves the field more than a quarter smaller than before the pandemic.

The concentration in senior roles creates its own problem. Experienced engineers were once junior engineers who learned by fixing bugs and writing basic features. If those tasks go to AI agents and employment for 22-to-25-year-olds keeps falling, no one in the sources has shown where the next generation of senior developers comes from. Indeed's next postings data will show whether the entry-level gap narrows or keeps widening.