Initial unemployment claims rose to 200,000 for the week ended May 2, according to Labor Department data released Thursday. That's a 10,000 increase from the prior week's revised figure of 190,000, but still below the 205,000 economists expected.

The number is notable for what it doesn't show: any sign that the wave of AI-justified layoffs hitting the technology sector is bleeding into the broader economy. Weekly claims have stayed below 230,000 all year, even as major tech employers have announced cuts totaling more than 120,000 workers in 2026 alone.

AI as the new restructuring rationale

The tech layoff tracker TrueUp counts 273 layoff events at tech companies so far this year, affecting 121,111 workers. A separate analysis from Nikkei Asia found that roughly 48 percent of first-quarter tech layoffs were explicitly attributed to AI and automation.

Meta announced in April that it would eliminate 10 percent of its workforce, roughly 8,000 people, with cuts beginning May 20. Coinbase cut 700 workers, framing the move as part of becoming "lean, fast, and AI-native." Block slashed its headcount nearly in half. Oracle has quietly shed more than 10,000 positions.

These are real job losses, and they are hitting some workers hard. But the pattern that's emerging is a sectoral reshuffling, not a macroeconomic contraction.

The gap between announcements and claims

Why aren't tech layoffs showing up more clearly in jobless claims? A few factors help explain it.

First, generous severance packages common at major tech employers delay when workers file for unemployment. Economists speculate that laid-off technology workers are most likely receiving those packages, which can run several months before benefits eligibility kicks in.

Second, many tech companies are conducting what amounts to a labor reallocation. Bloomberg data suggests roughly half of AI-attributed layoffs will result in the same roles being rehired offshore or at lower salaries. Customer support, quality assurance, and content moderation roles are disappearing. Machine learning engineers and AI infrastructure specialists are in shortage. The people losing jobs are generally not the same people getting hired into the new ones.

Third, the rest of the economy is holding up. The Bureau of Labor Statistics reported 6.9 million job openings in March, translating to 0.95 openings per unemployed person, up from 0.91 in February. That ratio remains consistent with a labor market in rough balance rather than deterioration.

Sentiment is steady for now

Consumer perception of the job market is stable. According to The Conference Board's April consumer confidence survey, 27.3 percent of respondents said jobs were "plentiful," virtually unchanged from March. Meanwhile, the share saying jobs were "hard to get" actually declined, falling to 19.8 percent from 21.3 percent.

On the employer side, the Conference Board's CHRO Confidence Index hit a new high of 59 in the first quarter, with 59 percent of chief human resources officers saying they expect to increase hiring in the first half of the year.

That doesn't mean the AI transition will be painless. The workers being displaced often have skills that don't transfer easily into the roles now in demand. The Information sector saw job openings decline 33 percent year over year in March, the steepest drop of any private industry. For technologists whose roles have been automated away, landing a new position is taking longer.

The numbers to watch

Economists forecast that nonfarm payrolls increased by 62,000 in April, a slowdown from March's 178,000 gain but still above the break-even rate needed to keep pace with working-age population growth. The unemployment rate is expected to hold around 4.3 percent, possibly rounding down to 4.2 percent.

The labor market remains remarkably resistant to the anxiety that tech layoff headlines might otherwise generate. That resistance has limits. If hiring slows further or if severance packages at tech firms start running out in volume, the claims data could shift. For now, the trend line says stability.