AI Is Cutting Jobs Everywhere Except Engineering. Here's the Data.

# AI Is Cutting Jobs Everywhere Except Engineering. Here's the Data.

Published: July 16, 2026

Two numbers landed within weeks of each other this summer, and they seem to contradict each other. As of July 15, 2026, there have been 302 layoff events in 2026 impacting 201,754 workers, with employers citing AI as the primary driver behind almost 40% of May's announced cuts. (CBS News, CNBC)

At the same time, SignalFire's 2026 "State of Talent Report" found that engineers made up 55% of all new hires in 2025 across the twelve companies it classifies as Tech Majors — Alphabet, Meta, Apple, Amazon, Microsoft, Netflix, Nvidia, Tesla, Uber, Airbnb, Block, and Stripe — up from just 46% in 2019. While total hiring at those companies fell 25% versus 2019 levels, engineering hiring fell only 11%. (TechCrunch)

Both numbers are real. They're not describing the same job market — they're describing two different halves of it, and understanding the split is the single most useful thing an AI engineer can do for their career planning right now.

The Layoffs Are Real, But They're Not Where You Think

Oracle alone has cut an estimated 21,000 jobs in 2026 attributable to AI-driven restructuring, with more reportedly on the way. (Forbes) Roughly 120,000 tech roles have been cut in 2026 overall, and AI-driven layoffs have spread well beyond tech — into finance, logistics, consulting, media, retail, and manufacturing. (CBS News)

But the CBS News and CNBC reporting both flag the same pattern underneath the topline numbers: companies are simultaneously laying off in some departments while hiring aggressively in others. The functions getting cut are recruiting, marketing, customer support, and experimental or duplicated projects. The functions getting built out are AI, security, and core product development — with machine learning infrastructure, model evaluation, AI safety, and applied research explicitly called out as roles in acute shortage even as the overall headcount numbers shrink. (CBS News)

In other words: the "AI layoffs" headline number is mostly a story about AI *replacing the work AI is good at* — repetitive analysis, first-draft content, tier-1 support, templated outreach — not a story about AI replacing the people who build AI systems.

Why Engineering Is the Exception

SignalFire's report, which tracked the careers of millions of employees across more than 80 million companies, is the clearest data point yet that engineering has decoupled from the broader hiring slowdown. Two findings stand out:

1. Early-stage startups hired 7% more engineers in 2025 than in 2019 — meaning AI tooling isn't just failing to shrink engineering demand at new companies, it's *increasing* it. The read: AI coding assistants let a smaller team ship more, which makes each additional engineering hire more leveraged, not less necessary. (TechCrunch)

2. The 55% new-hire share isn't evenly distributed across engineering — it's concentrated in roles adjacent to AI systems themselves. This lines up with what LLMHire's own listing data shows: LLM fine-tuning is now the single hardest AI specialization to hire for, with demand up 135.8% in 2026, and agentic AI job postings grew 280% year-over-year to roughly 90,000 US listings. (KORE1, Second Talent)

Put those together and the picture sharpens: the resilience isn't "engineering" as a broad category, it's specifically the engineers building, evaluating, securing, and fine-tuning the AI systems that are simultaneously enabling the layoffs happening around them.

The Skills Gap Behind Both Numbers

The demand side backs this up at scale. Industry estimates now put unfilled AI engineering roles at roughly 1.6 million globally against fewer than 518,000 qualified candidates — about 3.4 open roles per qualified engineer. (Qureos, KORE1) LLM specialists with production fine-tuning experience (LoRA, QLoRA, RLHF) are commanding $220K–$280K base, with fine-tuning skills alone adding another 10–15% premium at mid-level. (Second Talent)

That shortage is the mechanism connecting both headlines. Companies can cut headcount in functions where AI tooling has closed the skills gap (support, first-pass content, routine analysis) while being unable to cut — and in most cases still desperately hiring for — functions where the skills gap is still wide open (fine-tuning, evaluation, agent orchestration, AI safety). The layoffs and the hiring surge aren't contradictory. They're the same underlying force pointed in opposite directions depending on how replaceable the work is.

What This Means If You're Job Hunting Right Now

• "AI is cutting jobs" is not a reason to avoid AI engineering roles — the roles being cut and the roles in shortage are different roles, often at the same company. Read layoff announcements for *which function* was cut, not just the headline count.
• Generalist software engineering is safer than non-engineering roles, but AI-adjacent engineering is safer still. If you can move your resume toward fine-tuning, evaluation, or agent/orchestration work, the data says that's where the 3.4-to-1 shortage ratio is most favorable to candidates.
• Early-stage startups are quietly a strong bet. The 7% year-over-year increase in startup engineering hiring (against a shrinking baseline everywhere else) suggests smaller companies are using AI leverage to justify *more* engineering headcount, not less.
• Don't confuse "engineering is resilient" with "every engineering job is safe." Roles that are closest to templated, high-volume, low-judgment work are the ones both AI tooling and AI-driven restructuring target first — inside engineering orgs too, not just outside them.

Where to Find These Roles

LLMHire tracks AI engineering roles across LLM fine-tuning, agent orchestration, ML evaluation, and AI safety — the exact specializations showing up as both hardest-to-fill and best-compensated in the 2026 data above — sourced from Greenhouse, Lever, Ashby, and direct company listings, updated 6× daily.

Browse LLM fine-tuning and RAG roles →

See agentic AI and agent orchestration roles →

Explore AI safety and evaluation roles →

Related: 185,894 Layoffs, $285K AI Salaries: Inside the Widest Split the Tech Job Market Has Ever Seen · Meta's AI Layoffs: Where 500+ Displaced Engineers Are Landing in 2026 · LLM Engineer Salary Benchmarks 2026

LLMHire tracks 6,450+ AI engineering roles from Greenhouse, Lever, Ashby, and direct company listings. Updated 6× daily.