Did GOOGL's Jul 22 capex-guidance selloff break it from fellow hyperscalers?

Treated: GOOGL. Donors: MSFT, META, AMZN (fellow Mag-7 hyperscalers, all running their own large AI-capex programs and facing similar market scrutiny over it) — hand-picked, not algorithmic. Treatment date is GOOGL's Jul 22 after-close report: Cloud grew 63% YoY, a strong number, but $190B capex guidance crushed the FCF margin (21%→10, from 21%→9.2% per prior notes) and the stock sold off anyway. Tests whether this was a GOOGL-specific reaction to its own guidance, or part of a shared 'market re-pricing hyperscaler AI-capex risk' move that should also show up in the donor pool.

Actual vs. synthetic counterfactual

8091102114125 Treatment date Actual Synthetic
Actual Synthetic (counterfactual)

Fit quality

Fit-window RMSE4.5817
RMSE ÷ fit-window std dev0.46×
Fit-window days125
Confound-window mean gapn/a (fit-end = treatment date)
Post-period mean gap-3.47

Placebo test — same method run on every unit, treated one highlighted

UnitMean post-period gapMax abs. gap
MSFT-9.22%9.22%
AMZN+5.17%5.17%
META+4.55%4.55%
GOOGL (treated)-3.47%3.47%

Verdict

GOOGL's post-period gap ranks #4 of 4 units by absolute size in the placebo test (larger rank = more likely a real, unit-specific effect rather than noise the method would produce for any random unit).

Read: this does NOT look like a clearly unit-specific effect — other units in the placebo test show gaps of similar or larger size, so this could just be estimation noise rather than a real, distinct move.

Generated 2026-07-27 07:38 UTC · causal-toolkit · synthetic control (Abadie-Diamond-Hainmueller, outcome-only variant) · donor pool hand-picked, not algorithmically matched · treat every result here as directional, not a p-value.