120,000 Layoffs, AI Explicitly Cited: The Attribution Threshold Has Been Crossed
textak places the probability of a major AI-attributed layoff wave at 73%, and today's data doesn't just support that position — it arguably resolves it. As of early July 2026, 267 layoff events have explicitly cited AI as the primary cause, affecting 156,270 workers across 150 companies. The question was never whether displacement would happen. It was whether companies would say so publicly. They are saying so, loudly, while posting record revenues.
Our 73% reflected a specific structural bet: that investor pressure for demonstrated AI ROI would eventually override corporate instincts to avoid displacement optics. That bet is paying out. The TechCrunch and SkillSyncer data points to something harder to dismiss than a handful of anecdotes — 56% of layoff events in 2026 have explicitly named AI as the driver, averaging roughly 989 job losses per day. Goldman Sachs analysis puts the monthly figure at 16,000+ payroll cuts. This is not a pattern of quiet attrition with retroactive AI attribution. Companies are leading with the explanation.
The Microsoft announcement is illustrative of the mechanism we identified. Four thousand eight hundred roles eliminated, with the company simultaneously announcing expanded AI infrastructure investment and Copilot scaling partnerships across ASEAN. The framing is explicit: AI tools replace the coordination overhead of certain engineering and commercial roles. Microsoft isn't burying this. They're presenting it as strategic logic. Same structure at Meta: 8,000 employees cut, 7,000 repositioned into AI-focused roles. The ratio signals intentional restructuring, not cyclical cost-cutting.
Where we need to be honest about our model: the 73% was calibrated around 'first major wave publicly attributed to AI,' which implied a threshold event — a single, visible, undeniable attribution. What we got instead is a distributed pattern across 150 companies. That's arguably more significant than a single headline event, but it's worth asking whether the forecast resolves on distributed acknowledgment or requires a more concentrated signal. We're comfortable reading today's data as resolution-class evidence, but we're naming the ambiguity rather than papering over it.
The counterargument worth engaging: most displacement may still be attrition-based, with AI attribution functioning as convenient framing for restructuring decisions that would have happened regardless. The 'record revenues plus layoffs' pattern could reflect post-pandemic normalization rather than genuine AI substitution. We weight this less heavily than we did six months ago, because the role categories being cut — customer support, content moderation, QA testing, traditional software engineering — map precisely onto the tasks where AI capability is most mature. That's not coincidence. What would push us below 60%: a rigorous study showing that layoff rates in AI-exposed roles are statistically indistinguishable from non-AI-exposed roles controlling for sector and revenue cycle. We haven't seen that study, and the directional evidence we have points the other way.