Older Americans searching for work in April 2026 faced sharply longer spells of joblessness than their younger counterparts, with 34.5 percent of unemployed people aged 55 and older stuck without a paycheck for 27 weeks or more. That share ran roughly eight percentage points above the rate for workers under 55, based on age-stratified data drawn from the federal household survey that feeds the monthly jobs report. The gap is widening at the same moment that major employers are announcing AI-driven restructurings, raising pointed questions about which workers will absorb the displacement.
What is verified so far
The Bureau of Labor Statistics published its monthly employment situation report for April 2026 on May 8, 2026. Table A-12 of that release defines long-term unemployment as a duration of 27 weeks or more and provides the national share of unemployed workers who fall into that category. The 34.5 percent figure for the 55-and-older cohort, and the roughly eight-point spread over younger jobseekers, are derived from the same Current Population Survey (CPS) microdata that underpin the official release. Those microdata files for April 2026 are available through the Census Bureau’s CPS download page, which contains the age, labor-force status, and unemployment-duration variables needed to reproduce the cross-tabulation. The CPS methodology hub at BLS confirms the standard definitions: a person is classified as unemployed if they had no employment during the survey reference week, were available for work, and had made specific efforts to find a job in the prior four weeks.
These definitions matter because they separate long-term unemployment from short spells between jobs. A worker who left a job in March and is still looking in April is unemployed, but not yet long-term unemployed. By contrast, a 58-year-old who has been sending out résumés since last fall without success is counted among the long-term unemployed once the search stretches past the 27-week mark. The 34.5 percent figure indicates that more than one in three unemployed older workers were in that latter, more precarious category in April.
On the corporate side, Cloudflare filed a Form 8-K with the Securities and Exchange Commission on May 7, 2026, disclosing a restructuring plan tied to what the company called its evolution to an “agentic AI-first operating model.” The 8-K filing stated that Cloudflare expects to cut its workforce by approximately 20 percent, with estimated charges of $140 to $150 million. An accompanying exhibit pegged the reduction at roughly 1,100 people. Cloudflare projected that most of those charges would land in the second quarter of fiscal year 2026 and that the restructuring would be substantially complete by the end of the third quarter.
Cloudflare’s disclosure is part of a broader pattern in which firms explicitly link headcount reductions to automation or AI adoption. In this case, the company framed the cuts as necessary to reorient staffing and spending toward AI agents and related infrastructure. While the filing does not spell out which roles are being eliminated, the description suggests a shift away from more traditional operational and support functions toward positions that build, maintain, or deploy AI systems. For workers whose skills do not align with that pivot, the restructuring represents a sudden and potentially lasting detour from steady employment.
What remains uncertain
The standard published tables in the Employment Situation release do not pre-compute the exact 34.5 percent long-term unemployment rate for the 55-and-older group. That figure requires custom extraction from the CPS Basic Monthly microdata, cross-tabulating unemployment duration by age bracket. Independent researchers can replicate the calculation using the publicly available April 2026 files, but the BLS itself does not highlight this specific breakout in its summary tables. Until outside analysts or organizations such as AARP publish their own tabulations, the precise eight-point gap between older and younger workers carries a small margin of methodological variation depending on how age cutoffs are defined and whether the comparison group spans ages 16 to 54 or uses a narrower band.
There is also uncertainty around how persistent this gap will be. One month of data can reflect temporary shocks, seasonal patterns, or sampling noise. To know whether older workers are settling into a structurally higher risk of long-term unemployment, analysts will need to track several more months of CPS microdata. If the elevated share of long-term unemployed older workers remains in place or widens, that would point to deeper issues in how the labor market is absorbing demographic change and technological disruption.
No primary source directly connects AI-related layoffs to age-specific unemployment outcomes. Cloudflare’s SEC filing describes the size and rationale of its workforce reduction but does not disclose the age distribution of affected employees. Drawing a causal line from one company’s restructuring to the national long-term unemployment rate for older workers requires data that neither the 8-K nor the CPS release currently provides. WARN Act notices and future CPS releases could eventually fill that gap, but as of May 8, 2026, the link between AI-driven cuts and age-stratified displacement is circumstantial rather than documented.
The broader pattern of firms citing AI as a reason for headcount reductions is visible across earnings calls and press releases in early 2026, yet aggregate tracking of how many separated workers are over 55 does not exist in any single federal dataset. The Department of Labor collects layoff data through several programs, but none currently tags separations by the employer’s stated AI rationale and the worker’s age simultaneously. That means policymakers, advocates, and affected workers are left to infer the distributional impact of AI-related restructuring from partial indicators, case studies, and small-sample surveys rather than a comprehensive national series.
How to read the evidence
Two distinct categories of evidence anchor this story, and readers should weigh them differently. The CPS microdata are primary survey evidence collected by trained Census Bureau interviewers from roughly 60,000 households each month. When a statistic is computed directly from those files, it reflects a probability sample of the entire civilian noninstitutional population, with known sampling error. The long-term unemployment share for older workers falls into this category: it is a measurable, replicable statistic, not an estimate from a model or a projection.
Cloudflare’s SEC filing sits in a second tier of primary evidence. An 8-K is a legally required disclosure, and the company faces liability for material misstatements. The headcount reduction of approximately 1,100 people, the $140 to $150 million charge estimate, and the stated rationale of shifting to an agentic AI-first model are all on the regulatory record. What the filing does not provide is demographic detail on who is losing their jobs. That absence is itself informative: it highlights a gap between the precision with which companies describe technology strategy and the opacity surrounding which workers bear the cost.
For readers trying to make sense of the intersection between an aging workforce and rapid AI adoption, the safest conclusions are narrow. It is well supported that older unemployed workers in April 2026 were more likely than younger peers to be out of work for six months or longer. It is also documented that at least one large technology company is cutting roughly a fifth of its staff as it reorganizes around AI. What remains unknown is how much those trends overlap in practice-how many of the displaced workers at Cloudflare and similar firms are in the 55-and-older bracket, how quickly they find new jobs, and whether they eventually show up in the long-term unemployment statistics.
Until more granular data arrive, the evidence points to a labor market where age and technology change are interacting in ways that are measurable at the aggregate level but murky in their mechanisms. Older workers already face longer jobless spells, and AI-driven restructurings are adding fresh waves of displacement. Whether policy responses and employer practices can close that gap-or whether it will widen further as AI tools spread-will be one of the central employment questions of the coming years.
