Cross-cutting analysis
The stock pages underwrite one name at a time. These pieces read across the universe — cohorts, category-wide narratives, and the places where the market's story and the framework's scores disagree. Every score cited is computed live from the same data, so nothing here goes stale quietly.
Meta's Capex Has One Receipt, and It Is Price Per Ad
Meta spent $31.1B in a quarter and printed $784M of free cash flow, and the tape traded the capex line. The line that tests the build is the other one: price per ad up 12% for a second straight quarter while impressions decelerated. Only the outcome owner can charge for a better match.
ServiceNow and the Seat-Pricing Question
The market spent the first half of 2026 pricing enterprise software for the collapse of per-seat billing. ServiceNow's Q2 is the first hard read on that fear — and it suggests the market indexed on the wrong variable.
S&P Global and the Half With the Moat
The market marked S&P Global down on a comparison with a company that no longer exists. The print underneath says something more useful: in financial infrastructure the growth sits in the half of the business that owns a benchmark, not the half that sells data.
The Odds on Alphabet's $200B AI Build
Alphabet beat on revenue, on Cloud and on margin, then fell about 5% because capex went up. The market is sorting hyperscalers by how much they spend. The variable that decides whether a build pays off is how much of it was sold before it was poured.
Numbers stay live
Articles carry tickers, never scores. Every moat, growth and valuation figure resolves against the coverage registry at render time, with valuation recomputed from the current price.
Every thesis is falsifiable
Each piece states, in one sentence, the observation that would prove it wrong — and carries the date it was last checked against the underlying data.
Cross-cutting by design
A stock page underwrites one name. These read across four or more, where the comparison itself is the argument and the market has a story the data disputes.
Every piece has a Markdown mirror at /research/[slug]/llms.txt for agents and readers who prefer plain text. Start from the coverage universe to see the underwriting each argument is built on.