For the past three years, the dominant narrative has been that AI will hollow out software engineering. Coding assistants, autonomous agents, and “vibe coding” tools would replace the people writing the code. New data from venture firm SignalFire puts a serious dent in that story.

According to SignalFire’s 2025 State of Tech Talent Report, which tracked career movements across more than 80 million workers and thousands of companies, engineering is the single most resilient job function in the tech sector right now. Not sales, not finance, not operations. Engineering.

What the Numbers Actually Show

Total hiring across large technology companies dropped roughly 25% in 2025 compared to 2019 levels. Engineering hiring fell too, but only 11%. When you look at the share of new hires who are engineers, it went up, not down. Engineers made up 55% of all new hires in 2025 at the twelve companies SignalFire classifies as “Tech Majors” (Alphabet, Meta, Apple, Amazon, Microsoft, Netflix, Nvidia, Tesla, Uber, Airbnb, Block, and Stripe). In 2019, that share was 46%.

At early-stage startups, the picture is even clearer. These companies collectively hired 7% more engineers in 2025 than they did in 2019, a period that includes the AI boom and all the uncertainty that came with it.

The companies building the most powerful AI tools in the world, Anthropic, OpenAI, and Meta, are growing their engineering teams two to three times faster than they are losing engineers. Anthropic in particular posted an 80% two-year retention rate, which is exceptional in an industry where eighteen-month job-hops are the norm.

Where the Pain Is Real

The data is not uniformly positive. Entry-level roles have taken a genuine hit. The share of recent graduates landing jobs at the Magnificent Seven tech companies (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla) has dropped by more than half since 2022. New grads competing for their first technical job are walking into a harder market than their predecessors did.

The job market hasn’t collapsed, but it has bifurcated. Experienced engineers, especially those who can work alongside AI tools effectively, are in higher demand than ever. Junior roles that were once a reliable pipeline into the industry are fewer and more competitive.

Why Engineering Grew Relative to Everything Else

The answer is not that AI failed to change engineering work. It changed it significantly. The answer is that AI made engineering more central to every business, not less.

More companies are now software companies. More operations depend on AI-driven systems that need to be built, maintained, integrated, and iterated on. Every deployment of an AI agent, every automation of a business process, every custom AI application requires someone who can architect it properly, connect it to real data, and stop it from doing something unintended.

The tools got more powerful. The demand for people who know how to use them got bigger. That is the dynamic the displacement narrative missed.

What This Means for Business

If you are a business owner or executive thinking about AI strategy, this data should reshape how you see the talent question. The answer to “how do we use AI” is almost never “hire fewer technical people.” It is usually “hire or develop people who can put AI to work.”

The companies pulling ahead right now are the ones investing in technical capability, either building it in-house or accessing it through partners. The companies falling behind are the ones treating AI as a cost-cutting mechanism applied to human headcount before they have figured out what they actually want AI to do.

If you are a data professional, analyst, or engineer, the job displacement headlines are not the full picture. Your skills are more valuable in an AI-augmented environment than in an environment without it, because you can apply AI effectively in ways that non-technical colleagues cannot. The challenge is staying current, not becoming irrelevant.

If you are managing a team and wondering whether to invest in upskilling versus replacing roles with AI, the evidence from the most AI-intensive companies on earth suggests upskilling is the right bet. Anthropic’s 80% two-year retention rate does not happen by accident. It reflects how much experienced technical talent is worth when it is oriented around AI-native work.

The Entry-Level Caveat

It would be wrong to ignore the entry-level squeeze. The traditional path of taking a junior developer or junior analyst role at a big company to build foundational skills is narrower than it was. Companies are hiring experienced people and expecting them to ramp quickly with AI assistance, rather than building large cohorts of graduates who learn on the job over two to three years.

This means the way people develop entry-level skills needs to change. Online learning, structured courses, applied projects, and certifications have to do more of the work that structured graduate programs used to do. It also means the people who invest in that structured learning before entering the job market will have a real edge.

The Bottom Line

Engineering jobs are not disappearing in the AI era. They are concentrating around the people most capable of making AI work, which means technical skills are worth more, not less. The displacement risk is not for engineering as a function. It is for individuals who do not adapt their skills to work alongside AI tools.

The gap between “AI-augmented engineer” and “traditional engineer” is getting wider every month. That is where the real career risk lives, and it is also where the opportunity is.

---

Enterprise DNA helps data professionals and business teams stay ahead of that gap. Whether you are building your own skills or equipping your team, explore the EDNA Learn platform or talk to us about business upskilling programs.