The narrative was simple. AI would write the code, the junior engineers would be sidelined, and the headcount would shrink. Silicon Valley spent the last eighteen months bracing for a mass extinction event in software development.

It hasn't happened.

While tech layoffs hit a multi-year high in May, the underlying reality of the labor market tells a different story. New research from the venture firm SignalFire suggests that engineering is not just surviving; it is the most resilient job function in the industry. The machines are not replacing the builders. They are simply changing what it means to build.

The Hiring Reality vs. The Layoff Narrative

Layoff announcements are often performative. Companies cite AI as a convenient scapegoat for broader restructuring, claiming one engineer can now do the work of five. But tracking layoffs is notoriously unreliable. People delay updating their LinkedIn profiles. Severance packages mask the true timeline of departures.

SignalFire looked at the only metric that matters: hiring. By tracking millions of careers across 80 million companies, they found that engineering roles are holding steady. While total hiring at large tech firms dropped 25 percent since 2019, engineering roles saw a decline of just 11 percent.

At the twelve companies classified as "Tech Majors"—including Meta, Nvidia, and Stripe—engineers accounted for 55 percent of all new hires in 2025. That is a sharp increase from 46 percent in 2019. If AI were truly a substitute for human talent, these companies would be the first to cut. They aren't. They are doubling down.

The Jevons Paradox in Code

Why is the demand for engineers rising as coding becomes easier? The answer lies in the Jevons paradox. In economics, this occurs when increased efficiency leads to higher consumption, not lower.

When you make a resource cheaper, you don't use less of it. You find more ways to use it.

Nvidia CEO Jensen Huang has been the most vocal proponent of this view. He argues that agentic AI has made his engineers busier than ever. The AI writes the boilerplate code in seconds. The engineer then spends that saved time architecting the next complex feature. The work expands to fill the new capacity.

"They’re suddenly a lot more productive," says Asher Bantock, SignalFire’s head of research. "And there’s endless work for them to do."

What This Means for Developers

For the individual developer, the job description is shifting. The era of manual syntax grinding is ending. The era of systems thinking has begun.

Entry-level roles are the most exposed, but not because they are being deleted. They are being elevated. The barrier to entry for building a functional application has collapsed. A junior engineer today can ship a product that would have required a senior team five years ago.

Even at Anthropic, a company at the bleeding edge of AI development, the data remains quiet. Peter McCrory, the firm's head of economics, noted that there is no material difference in unemployment rates between jobs heavily exposed to AI and those that require physical interaction. The "AI-driven job apocalypse" is not showing up in the labor statistics.

Key Takeaways

  • Hiring is the true signal: While layoffs grab headlines, engineering hiring remains significantly more resilient than other corporate functions.
  • Efficiency creates demand: AI tools are triggering a Jevons paradox, where increased coding speed leads to more ambitious projects rather than fewer engineers.
  • The shift is structural: Tech majors are allocating a higher percentage of their total headcount to engineering today than they were in 2019.

The Next Frontier

The next six months will be the real test. As companies move from experimental AI pilots to full-scale production, the pressure to prove ROI will intensify. If engineering headcount remains stable through the next round of quarterly earnings, the debate will effectively be over. The question will no longer be whether AI replaces engineers. It will be how much more we can build with the engineers we have.