In May 2025, we looked at two weeks of hiring activity, a batch of new postings from Google, a wave of new roles at Microsoft even after a 7,000-person layoff, and called it evidence that tech was reshuffling rather than collapsing. That read was built on a snapshot. Fifteen months on, with a full year of layoff and hiring data behind it, it is worth checking that call against the aggregate record rather than another two-week window.
The honest answer is mixed. Some of the reshuffle framing held up. Some of it did not.
What the layoff data actually shows
US tech layoffs did not disappear after 2025, and by mid-2026 they were accelerating again. According to Crunchbase News, which tracks layoffs.fyi data, US tech companies cut at least 127,000 workers in 2025, up from roughly 95,700 in 2024. That alone complicates the idea that tech had turned a corner.
2026 has been worse in aggregate, not better. Challenger, Gray & Christmas, the outplacement firm whose monthly job-cut reports are the standard industry benchmark, recorded 139,156 tech-sector layoffs in the first half of 2026, an 83% increase over the 76,214 tech cuts in the first half of 2025. Tech accounted for nearly a third of all US layoffs announced in that period, out of 443,604 total layoffs across all industries (Challenger, Gray & Christmas, H1 2026 report, reported by HR Dive, July 2, 2026). Of the layoffs announced through June 2026, Challenger attributed 101,743, about 23% of the total, directly to AI-driven restructuring.
The most recent monthly data adds a wrinkle. Challenger's July 2026 report put total US layoffs at 33,429, the lowest monthly figure in two years and a 46% drop from July 2025. But tech moved the opposite direction: tech-sector cuts rose to 9,867 for the month, up 67% year over year (Challenger, Gray & Christmas, July 2026 report, reported by HR Dive, August 10, 2026). While layoffs across the broader US economy were cooling, tech layoffs were still climbing. That is not consistent with "tech is fine now."
What the hiring numbers show
The counterweight is that tech hiring kept moving through the same period, which is the part of the original call that held up better. CompTIA's analysis of BLS and Lightcast data for March 2026 counted more than 537,000 active tech job postings, up 9.7% from February and 8.9% year over year, with 254,000 new postings added that month alone (CompTIA, "Uncertainty and Caution Apparent in New Tech Employment Data," April 3, 2026). The unemployment rate for tech occupations that month was 3.9%, below the 4.1% national rate the Bureau of Labor Statistics reported for July 2026 (BLS, Employment Situation Summary, released August 7, 2026). Tech workers, in aggregate, are still finding new roles faster than the general workforce.
At the same time, CompTIA's own count showed tech occupation employment falling by 118,000 in March 2026 alone, and the BLS July report described the broader "information" sector, which includes software, data processing, and telecom, as essentially flat month over month. Flat headline employment with elevated layoffs and elevated postings running at the same time is not stability. It is churn: roles being cut in one place and opened in another, often inside the same set of companies.
Where Glozo's own data adds detail
Aggregate labor statistics show that churn exists. They do not show where it is concentrated. Glozo's own market-report data, drawn from live US job postings and candidate pools, points to a split by function rather than a uniform trend across "tech."
Two specialized technical categories in Glozo's August 2026 market reports are the tightest hiring markets currently measured on the platform. For cloud and infrastructure roles, the data shows 92,250 candidates in the pool against 6,761 active postings, about 14 available professionals per open role, described in the report itself as "the tightest market we have measured." For cybersecurity, the ratio is similar: 119,998 candidates against 7,865 active postings, roughly 15 per opening. Compare that to recruiting roles, a function outside the specialized technical core, where Glozo's July 2026 report found 327,487 candidates against 6,836 active postings, a 47.9-to-1 ratio, a deep-supply market by comparison.
| Function (Glozo market report, 2026) | Candidates per open role | Market read |
|---|---|---|
| Cloud & infrastructure | ~14 to 1 | Tightest measured |
| Cybersecurity | ~15 to 1 | Very tight |
| Recruiting | ~48 to 1 | Deep supply |
Put next to the layoff numbers, this is the clearest picture available of what "reshuffle" actually means in practice. The layoffs Challenger and layoffs.fyi are counting are real and are concentrated in generalist, support, and operations-adjacent roles, the ones most exposed to AI-driven restructuring and cost-cutting. Specialized technical disciplines, at least the two Glozo tracks closely, are not seeing that same contraction in hiring demand. They are competing for a shrinking, not growing, pool of qualified people even as headline layoff totals climb.
Three of the individual events feeding this macro trend already have their own coverage on Glozo. Block's February 2026 workforce reduction, in which the company attributed roughly 40% of cuts to AI-driven changes, is covered in What Block's 40% AI layoff tells every recruiter about 2026. Oracle's roughly 30,000-role reduction, and the contractor-rehire pattern that followed it, is covered in Tech layoffs are turning into contractor pipelines. Snap's April 2026 layoffs, and the gap between the roles it labeled "AI-driven" and the roles actually cut, are covered in Snap called these AI layoffs. Look at who they actually fired. Readers looking for the mechanics of any single company's cuts should start there; this piece stays at the industry level.
The honest verdict, 15 months later
The May 2025 call that tech was reshuffling rather than dying was directionally right about one thing: aggregate tech employment did not fall off a cliff, hiring never stopped, and specialized technical roles remained genuinely hard to fill throughout the period. It was wrong, or at least incomplete, in implying the reshuffle was painless or evenly spread. It was not. Tech layoffs in 2025 exceeded 2024's total. Tech layoffs in the first half of 2026 ran 83% ahead of the first half of 2025, and by July 2026 tech was cutting jobs even as the rest of the US economy was cutting fewer. Nearly a quarter of the cuts announced through mid-2026 were tied directly to AI restructuring.
So the more accurate framing, a year and change later, is not that tech is collapsing or that it is fine. It is that the industry is contracting in specific functions, mainly the ones easiest to automate or consolidate, while specialized technical hiring stays tight enough that qualified candidates in categories like cloud infrastructure and security are still scarce. Both things are true in the same labor market at the same time.
What this means for recruiters right now
The practical takeaway has not changed much since May 2025, even if the data underneath it has: a company-wide layoff headline says very little about whether a specific role or function at that company, or in that industry, is easy or hard to fill. The number worth watching is the candidate-to-opening ratio for the specific function being hired, not the topline layoff count for "tech" as a category. Recruiters working specialized technical roles should expect the market to stay tight regardless of what the layoff totals say, and should treat industry-wide headlines as context, not as a signal about any individual req.