On 8 Oct 2026, reports that OpenAI's annualized revenue was about $50B (not the ~$70B that had been circulating) hit AI stocks. If the market was pricing OpenAI risk, the companies most tied to OpenAI should have fallen hardest. Did they?
fell more on 8 Oct than the without one. By how many percentage points?
And which explained more of who fell?
Our stated prior (what we'd expect a smart reader to guess): OpenAI-linked names drop ~2x as hard, 3–4 pts more."Return beyond beta" = actual 8 Oct return minus beta × SOXX's 8 Oct return (). It strips out the part of the drop any equally jumpy chip stock would have had.
Mean 8 Oct return by layer (grey tick = beyond-beta). Layer groupings are editorial. Look at n: layers have only one or two names.
Tap a name. Labels come from a separate review pass: CONFIRMED = filing / IR release or 2+ independent outlets; UNVERIFIED = single non-primary source or not checkable. Items with no source found are counted, not shown.
Return = close 8 Oct ÷ close 7 Oct − 1. Beta = OLS slope of daily returns on SOXX daily returns, 60 sessions ending 7 Oct 2026. Correlations are Pearson. p-values: 10,000 random reshuffles of the OpenAI-link label (two-sided). Benchmarks on 8 Oct: .
Raw data and a script that recomputes every number on this page ship with the submission.