Remember when everyone said AI would turn engineers into obsolete code monkeys? That was the narrative. Tech layoffs were hitting hard, and the robots were coming for the white-collar jobs first. Software engineers were supposed to be the canaries in the coal mine.
Turns out, the canaries are thriving.
New data from SignalFire, a venture capital firm that tracks the tech talent market, tells a different story. While AI has dominated layoff headlines for the past two years, engineers are actually making up a larger share of new hires across the economy. Not just in tech — everywhere.
The numbers don't lie
SignalFire's data — pulled from millions of employee records and job postings — shows that engineering roles now account for 12.4% of all new hires in 2026. That's up from 9.8% in 2022, before the AI panic really set in.
Meanwhile, non-engineering roles? Their share of new hires has been flat or declining. Marketing, customer support, even design — all down as a percentage of new hires since 2023.
This isn't a blip. It's a structural shift.
“AI is not reducing demand for engineers. It's reducing the cost of non-engineering labor faster than it's reducing the cost of engineering labor.” — SignalFire analyst
The logic is brutally simple: AI can write marketing copy, generate images, and answer customer emails. It's getting decent at those things. But it still can't build complex systems, debug production issues at 3 a.m., or design scalable architectures that actually work. Those tasks require the kind of messy, contextual understanding that LLMs fumble.
Why engineers are safe
Think about what an engineer actually does. It's not just typing code — that's the easy part. Engineers spend most of their time figuring out what to build, how to integrate systems, and why something broke. They navigate legacy spaghetti code, negotiate with product managers, and translate business requirements into technical reality.
AI can generate a Python function in seconds. It cannot do any of the above.
This is why companies are hoarding engineering talent. They're realizing that AI tools are only as good as the people who wield them. You need engineers to prompt the AI, vet its output, and stitch it all together. Without engineers, AI-generated code is just a pile of plausible-looking garbage.
The data backs this up. According to SignalFire, the median engineering salary has actually risen 6% since 2023, adjusted for inflation. Meanwhile, salaries for roles like content writers and customer support agents have stagnated or dipped.
The layoff paradox
But wait — didn't we just see massive tech layoffs? Google, Meta, Microsoft — they all cut thousands of jobs. How can engineers be thriving?
Here's the nuance: those layoffs hit non-engineering roles disproportionately hard. At Meta, for example, the 2023 layoffs cut 30% of recruiting staff, 25% of marketing, but only 10% of engineering. Amazon's latest round in 2025 eliminated entire HR teams while leaving engineering headcount largely untouched.
Companies are trimming fat — and they've decided that engineers are muscle, not fat.
“If you're a company that wants to integrate AI into your product, you need more engineers, not fewer. The layoffs are a redistribution, not a reduction.”
Even the AI companies themselves are hiring engineers at a furious pace. OpenAI, Anthropic, and Google DeepMind have collectively added over 5,000 engineering roles in the past 18 months. They need people to build the next generation of models — and to clean up the mess the current ones make.
What this means for the rest of us
If you're an engineer, this is your moment. The data says you have leverage. You can ask for more money, more remote flexibility, more interesting projects. The market is on your side.
If you're not an engineer, the picture is murkier. The jobs that are most vulnerable to AI are the ones that involve pattern recognition, language generation, and routine data processing — exactly the things AI is getting competent at. That doesn't mean those jobs disappear overnight, but it does mean fewer new roles and slower wage growth.
The takeaway isn't that everyone should learn to code. That ship has sailed, and the market for junior engineers is actually saturated. The takeaway is that deep expertise — the kind that takes years to build — is becoming more valuable, not less. AI can mimic the surface level, but it can't replace actual depth.
So no, AI didn't kill engineering jobs. It made them bulletproof. The robots aren't coming for the builders — they're coming for everyone else.
