Meta and Microsoft Made the Trade-Off Visible. That's the Signal We've Been Waiting For.

First, the forecast definition. We've tightened this deliberately because the 'first' framing is indefensible without it. IBM's 2023 announcement pausing hiring for ~7,800 roles due to AI is real. Chegg's CEO directly linked student loss to ChatGPT before laying off staff. Duolingo cut contractors with explicit AI rationale in 2023. Block's Jack Dorsey just cut 4,000 jobs while stating AI could do the work 'more honestly' than humans — that's a direct quote, not an inference. If the forecast were simply 'has any company ever publicly attributed layoffs to AI,' it would have resolved YES eighteen months ago. The forecast we're tracking is narrower and more consequential: when does a company large enough to move institutional norms — Fortune 100 scale, SEC-filing language, named business unit — formally put AI displacement on the record in a way that creates legal and regulatory precedent? That threshold has not yet been crossed, and the distinction matters enormously for how employment law, labor contracts, and WARN Act obligations evolve.

So what did Meta and Microsoft actually say? Meta announced ~8,000 cuts affecting 10% of staff while simultaneously committing $135B in AI capex. Microsoft offered early retirement packages to 7% of its US workforce. Both companies have framed these moves in AI context — the trade-off between human headcount and AI infrastructure investment is structurally visible in their capital allocation. But 'structurally visible' is not 'formally stated.' We have not yet seen a Meta or Microsoft SEC filing that reads: 'We eliminated X roles in our [named unit] because AI systems now perform those functions.' The Dorsey quote at Block is the closest thing in today's news cycle to direct attribution — explicit, CEO-level, on the record — but Block has a ~$40B market cap. It's a proof of concept for the attribution behavior, not the institutional norm-setter our forecast requires. What today's news gives us is strong proximate evidence: the largest companies in the world are publicly performing the logic of the trade-off. The gap between performing the logic and filing the language is narrowing. That's why we moved from 67% to 70%.

The McKinsey data point deserves honest treatment. The claim that 1-in-4 entry-level job postings now require AI fluency versus 1-in-20 two years ago is evidence of skill-shift demand — not displacement volume. A company requiring AI fluency in new hires is equally consistent with augmentation (same headcount, different profile) as with replacement. We're not using it to prove displacement is happening. We're noting it as a lagging signal that the institutional expectation of human-AI substitutability is becoming mainstream enough to appear in job descriptions, which makes formal attribution psychologically easier for executives who follow. The S&P 500 headcount decline of 400,000 in 2025 — the first such decline since 2016 — is similarly proximate, not direct. Post-COVID overhiring corrections and interest-rate-driven efficiency mandates explain a substantial portion of that number. We cite it as scale context, not as proof that AI caused it.

What would move us? A Fortune 100 earnings call transcript or 8-K filing that explicitly names AI as the cause of FTE reduction in a specific unit would push us above 80% within weeks. Conversely, if Q2 earnings season passes with Meta, Microsoft, Google, and Amazon all continuing to use euphemistic 'efficiency' language without naming AI as the displacement mechanism — and if legal counsel continues to successfully suppress explicit attribution — we'd pull back toward 60%. The resolution window is calendar year 2026. The honest uncertainty in our model is this: we may be systematically overweighting CEO communication boldness relative to legal department caution. Dorsey can say what he said because Block's legal exposure is different from Meta's. The question is whether Zuckerberg's team will ever let him say it the same way.