"Are tech layoffs really driven by AI?"

"Yes, in large part. Hyperscalers and AI-exposed software companies are redirecting compensation budgets toward capex (servers, GPUs, data center buildout) and using AI productivity gains to justify smaller headcount. The result is contraction in white-collar tech and expansion in the trades that build the infrastructure."

"Which trades are hiring most aggressively?"

"Electricians lead, followed by HVAC and refrigeration techs, pipefitters and plumbers, gas turbine and power generation workers, welders, ironworkers, and heavy equipment operators. The common thread is power, cooling, and shell construction for data centers."

"Is now a smart time to leave tech for the trades?"

"It depends on your situation, but the case for the trades has rarely been stronger. Apprenticeship paths typically pay while you train, do not require a four-year degree, and lead to work that is hard to offshore or automate. Many tradespeople also run their own businesses later."

"How long does it take to retrain into a trade?"

"Trade school programs typically run from a few months to two years depending on the trade. Registered apprenticeships are usually 4 to 5 years and pay throughout. Many career-changers stack a short trade school program with an apprenticeship to accelerate entry."

Tech Layoffs vs Trade Hiring in 2026: The Career Inversion

Updated June 8, 2026 | TradeCareerPath Editorial Team

In 2025 and 2026, two stories ran in parallel and almost no mainstream coverage connected them.

Story one: big tech announced rolling rounds of layoffs, often framed as efficiency or AI-driven productivity. Story two: every electrical, mechanical, and civil contractor working on data centers was begging for workers.

These are the same story.

The Money Is Moving From Payroll to Capex

The hyperscalers and the largest AI-exposed software companies are funding the infrastructure buildout out of cash flow. To keep margins intact while ramping capex, they are trimming headcount in functions where AI tools have improved per-employee output: customer support, content production, parts of engineering, and middle layers of management.

The numbers are public. Combined capex guidance across Microsoft, Alphabet, Meta, Amazon, and Oracle is north of $300 billion for both 2025 and 2026, with a meaningful share earmarked for U.S. data center construction. At the same time, layoff trackers logged tens of thousands of tech cuts through 2025 and into 2026.

The compensation that comes off the income statement does not vanish. A meaningful portion of it shows up as construction spend, which lands as wages on construction job sites.

The Trade Side of the Ledger

Public industry data tells the other half of the story. U.S. data center construction starts hit roughly $77.7 billion in 2025, up 190% year over year, with more than $88 billion expected to break ground in just the first half of 2026 (Equipment World).

Industry estimates put the temporary construction labor associated with the announced and under-construction U.S. data center pipeline at roughly 4.7 million worker-positions, with hundreds of thousands likely to go unfilled at peak.

That demand pulls hardest on:

Electricians (medium-voltage, switchgear, controls)
HVAC and refrigeration techs (chillers, liquid cooling)
Pipefitters, welders, and plumbers (cooling loops)
Gas turbine and power generation technicians (firm power)
Heavy equipment operators, ironworkers, and concrete crews (shell)

For a deeper look at any of these, see the trade-specific guides linked at the bottom.

Why the Trade Path Holds Up Well

Three things make this moment unusually friendly to the trades:

Apprenticeships pay while you train. Most registered apprenticeships pay a wage and increase pay each year as you progress.
The work is hard to offshore. A medium-voltage feeder in Northern Virginia has to be terminated by someone in Northern Virginia.
AI is a tailwind, not a threat. The tools that disrupted knowledge work do not run a torque wrench. AI may change how trade businesses estimate, schedule, and document work, but it does not replace the field labor.

Pay Comparison Caveats

Direct pay comparisons between trade and tech roles are noisy. Tech compensation skews heavily on equity, and trade compensation skews heavily on overtime, per-diem, and union benefits that do not always show up in headline wage numbers. The honest answer is that experienced tradespeople running their own crews or owning their own businesses are competitive with mid-career software engineers in many markets, especially after factoring in zero student debt for those who took the apprenticeship route.

For a head-to-head look, see Electrician vs Software Engineer Career Comparison.

How to Move Into a Trade

Pick a trade with structural demand. Electrical, HVAC, plumbing, welding, and instrumentation are high on the list.
Choose a route: trade school, registered apprenticeship, or a stack of both.
Consider relocation if your current region does not have an active data center pipeline. See States Where the AI Buildout Is Hiring Trades.
Add the certifications data center contractors look for: OSHA 30, NFPA 70E, EPA 608, BAS coursework, welding endorsements as relevant.

Popular Trade Programs

About this guide: Researched and written by the TradeCareerPath Editorial Team. Our editorial team researches and sources every trade school and career guide using federal labor and education data, including BLS OEWS and Employment Projections, DOL apprenticeship records, IPEDS, College Scorecard, and state licensing boards. We follow the editorial standards documented at /editorial-policy/.