Tools & Notes

Costco is one of the highest-quality compounders in retail — a membership annuity wrapped around a low-margin merchandise engine, with returns on operating capital in the mid-20s and revenue still growing high-single-digits at $290B of scale. Routed to the capital-light engine, the fundamental-strength score pins at its ceiling and stays there across every plausible growth path. The whole five-year outcome then rides the one input the business cannot control: the terminal P/FCF. At ~48x free cash flow — the 58th percentile of its own ten-year band — the entry multiple is not cheap, and the exit multiple is doing most of the work.

A drawdown despite Azure at 40% and a fourth straight beat isn't an earnings problem — it's a capex-return problem. OCF compounds while capex growth is the wildcard: if the buildout plateaus by FY27, FCF inflects convexly off a flat spending line; if the arms race continues, FCF stays trapped and the multiple keeps compressing. The Burger Lens tool models the two lines separately; the FY27 capex framing on the next earnings call is the checkpoint.

Ecosystem, services, and brand are a genuine moat — but the 'no Jobs-like product lately' instinct has a concrete current form: Apple just conceded it can't build a competitive frontier AI model in-house and is licensing Google's Gemini to power Siri instead. Meanwhile the services moat itself is under live, not hypothetical, regulatory attack in both the EU and the US. Status: monitor — thesis intact today, but several of its load-bearing pillars are mid-stress-test, not resolved.

The bear case on Adobe boils down to 'AI is going to eat Adobe' — a mood, not an argument. The Decay Lens splits that into four drivers you can check against the 10-K, and the valuation math shows how much of it the market has already priced in.