Tech Is Laying Off and Hiring at the Same Time. Here's What's Actually Happening.

The contradiction looks impossible on paper. More than 130,000 tech workers lost their jobs in the first seven months of 2025, according to FinalRoundAI's tracking of 434 layoff events. Over the same stretch, broad hiring climbed 25% year-over-year, per Challenger, Gray & Christmas data. Both numbers are accurate. Reconciling them requires accepting that what used to be one engineering labor market has quietly split into two — and that most hiring strategies were built for the version that no longer exists.

The Numbers Don't Contradict Each Other — They Describe Two Different Markets

Indeed Hiring Lab's analysis through mid-2025 puts overall US tech job postings roughly 36% below their February 2020 baseline — and the gap isn't closing. At the same time, machine learning engineer openings climbed approximately 59% above that same pre-pandemic mark. Those two figures come from the same labor market, measured at the same moment.

Segment	Trend vs. Feb 2020 Baseline
Overall US tech job postings	Down ~36%
Machine learning engineer openings	Up ~59%
Software devs aged 22–25 (employment)	Down ~20% from 2022 peak
Engineers 30+ in AI-adjacent roles	Up 6–12%
Broad tech hiring, YoY change	Up 25%

What the aggregate masks: companies aren't pausing. They're reallocating. Microsoft eliminated more than 23,000 positions across three rounds between May and July 2025, explicitly linking the restructuring to AI investment. Intuit cut nearly 4,800 employees while its CEO framed the decision as redirecting capital toward the company's highest AI priorities. Recruit Holdings — the parent of Indeed and Glassdoor — cut roughly 6% of its HR technology workforce and cited AI as reshaping how employers and job seekers interact. The same logic is driving cuts and new headcount inside the same organizations.

Indeed Hiring Lab's researchers have raised a more unsettling possibility: the earlier collapse in tech postings may not be cyclical at all. If AI is structurally absorbing the output once attributed to headcount, the rebound that hiring managers have been waiting for may simply not arrive for certain role types.

What's Being Eliminated vs. What's Being Built

Stanford research tracking employment by age cohort puts the sharpest edge on this divide. Software developer employment among workers aged 22 to 25 fell roughly 20% from its late-2022 peak by mid-2025. Engineers aged 30 and over in AI-exposed roles saw employment rise between 6% and 12% over the same window. The pattern is consistent: entry-level code generation, boilerplate, and routine testing — the territory AI tools absorb most efficiently — is where the contraction is deepest. Complex system architecture, integration work, and production AI deployment remain stubbornly resistant to automation.

The roles actively growing across the market — AI engineering, AI infrastructure, prompt engineering, AI product management — require a fundamentally different skill profile than the roles being eliminated, according to FinalRoundAI's layoff data. The retraining gap between what's being cut and what's being hired isn't a temporary mismatch; it's the defining structural problem of this labor market right now. Filling it requires more than posting a job description.

What Engineering Leaders Should Actually Do With This

The instinct to post a job description and wait is the instinct to fail in this market. CompTIA and Dice data from the first half of 2025 shows that positions requiring 10 or more years of experience surged 17% even as junior postings contracted. Engineers who combine deep domain expertise with genuine AI fluency are rarely browsing job boards — they're working, they're selective, and they move when someone finds them, not when they search.

For engineers navigating the dislocation: the data is directional. Experience alone isn't a shield, but experience paired with AI capability commands roughly an 18.7% salary premium — approximately $25,000 more annually — above non-AI peers at equivalent seniority levels, according to Robert Half and Levels.fyi data. Building that combination deliberately is the move the numbers support.

Ryzlink's forward-deployed engineering model is designed for exactly this gap — sourcing AI-fluent engineers and typically placing them within 48 to 72 hours, accountable for what ships rather than just filling a seat. If your team is navigating this bifurcated market and the usual sourcing channels aren't producing the right profiles, Ryzlink's advisors can help.

The Long-Run Case for Acting Now

A two-year dislocation doesn't erase a decade-long trajectory. TechTarget's analysis of Bureau of Labor Statistics projections finds US tech jobs are on course to grow at roughly twice the rate of the overall workforce over the next ten years. July 2025 job cuts of 33,429 — the lowest monthly total since July 2024, per Challenger data — suggest the most acute phase of AI-driven restructuring may be past its peak, even if the recomposition of roles continues.

The companies using this moment to build AI-fluent engineering capacity — through hiring, retraining, or staffing partnerships — will hold a structural advantage when the churn settles. The talent exists; it's just not concentrated where anyone has historically looked for it. The firms that figure that out first won't need to be told twice.