Numbers about AI-driven layoffs are only as useful as the methodology behind them. That point, documented in April 20 hub coverage on the data war over AI attribution, is back in focus today with a new count that diverges sharply from the established Challenger figure.
Here’s what the registry confirms: tech sector layoffs in Q1 2026 have been tracked at over 73,000 positions across at least 95 companies, per April 21 hub coverage. Some counts approach 78,500 for the full quarter, a higher figure consistent with Q1 finalization as more company announcements were logged. Both numbers describe the same event. The difference is completeness of the counting window, not a contradiction.
The attribution divergence is a different matter. Challenger Gray & Christmas, whose methodology involves parsing company-reported layoff rationale, cited AI as a factor in approximately 25% of March 2026 layoffs. A separate tracker, the source of which has not been confirmed as of this publication, reportedly counted approximately 48% of Q1 tech layoffs as AI- or automation-attributed, representing roughly 37,600 positions from a reported total of approximately 78,500.
The source identity and methodology for the 48% figure are not confirmed. This is a significant caveat. The underlying data provider matters: Layoffs.fyi tracks announcements but doesn’t classify attribution the same way Challenger does. A company that announces “restructuring for efficiency in an AI-first environment” might be counted differently by each methodology. Use the 48% figure as reported and directional, not as a confirmed statistic.
Why do the numbers diverge so widely? Challenger counts attributions companies explicitly report. Broader trackers often infer attribution from surrounding context, job category, industry timing, concurrent AI investment announcements. Neither approach is wrong. They’re measuring different things: one measures what companies say, the other measures what the data pattern suggests. Goldman Sachs research from April 29 adds another layer, workers who were displaced describe AI as a factor even when company communications don’t.
What the methodological gap reveals is that AI attribution in layoff data is a contested measurement problem, not a settled empirical one. A 25% figure and a 48% figure can coexist when the definitional question, “what counts as AI-attributed?”, doesn’t have a consensus answer.
What to watch: Source confirmation for the 48% tracker figure is the immediate priority. If the data provider is identified as a named tracker with a documented methodology, the gap between 25% and 48% becomes the basis for the comparative deep-dive this story warrants. The April 20 “data war” narrative holds: the disagreement over how to count AI’s role in layoffs is itself a policy and compliance signal for organizations evaluating disclosure standards.
The broader implication is practical. HR leaders and workforce planners using layoff attribution data for policy or compliance purposes should treat any single figure, 25%, 48%, or otherwise, as a methodology-specific estimate, not a ground truth. The right question isn’t “what’s the number?” It’s “what does this particular method capture, and does that match what I need to know?”