Coinbase and Freshworks Cross the Line: AI Attribution Is No Longer Corporate Taboo — But Our Forecast Hasn't Resolved Yet
TexTak's [white-collar-displacement] forecast sits at 70%, up from 67% — and today's evidence is the most direct we've seen since we opened this forecast. Coinbase's CEO explicitly tied 700 job cuts to AI agent deployment. Freshworks' CEO publicly stated AI writes over half the company's code while cutting 11% of headcount during a revenue-growth quarter. These are not euphemisms. But our forecast hasn't resolved, and here's why that's not a contradiction.
Let's be precise about what we're forecasting and why it hasn't closed. Our resolution criterion requires a Fortune 50 or S&P 500 company making explicit AI-attribution for a significant layoff event in an earnings call or equivalent public filing — not just a mid-tier tech firm's CEO letter. Coinbase is a $40B company. Freshworks is smaller still. Both are meaningful data points, but neither is the category-defining event our resolution threshold targets. That distinction matters: a pattern of attribution at the mid-cap level is strong evidence the threshold will eventually be crossed, but it is not the crossing itself. We're at 70%, not 95%, for exactly this reason.
What these two cases do establish is that the attribution barrier is cracking. Coinbase CEO Brian Armstrong's framing — 'this is a new way of working, and we need to leverage AI across every facet of our jobs' — is direct causal language in a public document, tied to a specific headcount number, in a non-distress context (Freshworks grew revenue 16% YoY). This is qualitatively different from the vague 'efficiency' language that dominated 2023-2024 layoff announcements. We weight this heavily because it suggests legal and PR teams at these companies have concluded that AI attribution is now a feature, not a liability — a signal to investors rather than a risk to manage away. That's the mechanism we've been waiting to see.
On Cognizant: we want to be clear about evidentiary grade here. Project Leap is reported by multiple outlets as linked to AI-driven restructuring, but Cognizant has not issued official AI-attribution language. This is proximate evidence — conditions consistent with our thesis, not confirmation of it. We're watching for official statements, but until they arrive, Cognizant belongs in a different column than Coinbase and Freshworks.
The counterargument we take most seriously isn't the PR-risk argument anymore — that argument is weakening in real time. The structural concern that keeps us from moving higher than 70% is whether explicit public attribution at Fortune 50 scale is realistically achievable given WARN Act litigation exposure, discrimination lawsuit risk, and severance negotiation incentives. When a company says 'we cut these jobs because of AI,' it creates a paper trail that plaintiffs' attorneys can use to argue the displacement was foreseeable and targeted. That legal exposure may not disappear just because mid-cap tech CEOs are getting comfortable with the language. The Fortune 50 GC's calculus is different. Until we see one of them cross that line — in an earnings call, attributing specific headcount reduction to AI substitution — we're holding at 70%, not resolving.
What would move us? A Fortune 50 company citing AI as the explicit cause of a headcount reduction in a Q2 earnings call pushes us above 80%. A Q3 in which only smaller companies continue the attribution pattern while large-cap firms revert to efficiency language drops us below 55%. The 273 tech layoffs affecting 121,000 workers YTD represent massive scale — but scale of displacement and willingness to name the cause publicly remain two different variables, and our forecast is about the latter.