Walmart: A 9% Drop and Still No Margin of Safety
Walmart reported fiscal Q2 earnings before the open on August 20, 2026, and the stock fell roughly 9% during the session — a drop from $114.30 (the prior close) to an intraday $103.83. That is one session’s damage, not a verdict. This piece walks the four gates — circle of competence, moat, management, price — and asks whether the drop changed anything material. Short answer: it did not.
Gate 1: Circle of Competence
Walmart sells goods to roughly 280 million customers a week through more than 10,900 stores in 19 countries and its own e-commerce operation. FY26 revenue was $713.163 billion, with about 2.1 million employees. The bulk of the profit model is product margin at a net margin of just 3.07% — a business that works through turnover, not markup. Layered on top are three higher-margin attachments: membership fees (Walmart+, Sam’s Club), marketplace and fulfillment fees, and advertising built on the company’s own transaction data (Walmart Connect). Segments are Walmart U.S., Sam’s Club U.S., and International.
Is it knowable? The demand side is close to an analytical statement rather than a bet: people will buy groceries in ten years, and much of it at the cheapest large retailer. The information layer is solid — the Q2 8-K (accession 0000104169-26-000145) provided enough to reconstruct the quarter. The boundary is the channel question: whether ten years from now purchases flow through Walmart’s interface at all. The company itself, in the same 8-K’s forward-looking statements, flags “development and deployment of artificial intelligence technologies” as a competitive pressure. Demand is in scope; interface attribution is not.
Gate 2: Moat
The moat is cost advantage, primary, backed by efficient scale — store density lowers last-mile costs, and replicating that density requires burning the same capital again (FY27 CapEx guidance around 4.0% of net sales, roughly $28 billion annualized). Brand, membership, and marketplace network effects are secondary layers.
The litmus test runs backwards here. Buffett’s test — can you raise prices without a prayer meeting — does not apply. The company ran over 11,000 rollbacks this quarter and the CFO said tariff refunds would go first into price investments. A cost-advantage moat expresses itself not as high prices but as selling cheaper than rivals and still profiting. So the right question is whether competitors lose money at Walmart’s prices — and the competitor unit-economics data needed to answer is outside this filing. Pricing power: undetermined.
Direction: also undetermined, because both kinds of evidence sit in the same document. Widening: global e-commerce +23%, U.S. +24%, advertising +38% (Connect ex-VIZIO +43%), marketplace net sales +50% or more, store-fulfilled delivery +40%, membership fees +17% globally — high-margin layers growing several times faster than the core. Narrowing: U.S. comp ex-fuel decelerated from 4.6% to 2.6%, with ticket growth falling from +3.1% to +1.1%. Notably, transactions held at +1.5%, flat year over year — fewer people did not show up; baskets got smaller. Customers leaving and customers spending less say different things about a moat.
The AI-era overlay splits the same way. The advertising and marketplace layer should be strengthened: physical inventory location determines delivery speed, and delivery-and-purchase data sharpens ad targeting — assets no model replicates. The interface layer may be eroded: if purchase agents compare price and inventory directly, brand and app convenience get reduced to an API comparison — though the cost advantage itself survives that reduction, since cheapest wins. And much of this quarter’s comp slowdown belongs in neither bucket: pharmacy deflation from new maximum-fair-price regulation and higher fuel costs are regulatory and mix events, not demand events.
Gate 3: Management
Capital allocation shows no destruction — no chain of top-of-market acquisitions, no dilution — but neither does it actively impress. Buybacks of 42.3 million shares for $5.1 billion in FY27 year-to-date imply an average price of $120.57, well above the current $103.83 and far above our intrinsic value estimate. Dividends rose 5.3% to $0.99 per share half-year. CapEx is rising (3.74% of sales in FY26, guided toward 4.0%), and the 6-month FCF decline to $5.529 billion from $6.943 billion came from investment, not operating deterioration — operating cash flow actually increased.
On candor, three observations, none decisive. The company disclosed that tariff refunds (cited as a 750bps net benefit to adjusted operating income growth, on marketing pages rather than in the 8-K) partly offset its +17.4% adjusted operating income growth — that is disclosure, though the key interpretive number sits outside the legal filing. And within the same 8-K sit two unadjusted headwinds — 80bps from health and wellness, 125bps from pharmacy deflation specifically — whose distinction materially changes interpretation. Overall: conditional pass, low conviction.
Gate 4: Price
Owner earnings for FY26 fall in a wide band: $9.454 billion if all CapEx is maintenance, up to $21.893 billion if maintenance CapEx equals D&A ($14.203 billion, back-calculated). The range is 2.3x wide — but the conclusion does not depend on closing it. At a market cap of $826.287 billion (7.958 billion shares at $103.83), owner-earnings yield is 1.144% to 2.650%. Against the 10-year Treasury at 4.690%, that is −3.5 to −2.0 percentage points at either boundary — a gap no daily bond move closes.
Applying 12x–15x multiples (FY27 guidance: net sales +4.0–5.0% cc, adjusted operating income +7.0–8.5% cc) gives intrinsic value of $14.26 to $41.27 per share. The 30% safety-margin buy range is $9.98 to $28.89. The market price sits at 251.6% to 728.3% of intrinsic value. Even at the pre-drop $114.30, the distance to the upper buy line was −74.7%. The 9.2% drop did not meaningfully close it.
One honest caveat: the current price would require a 37.7x multiple on upper-bound owner earnings to justify, which could be argued if advertising, marketplace, and membership come to dominate the profit mix and warrant a platform multiple rather than a retail one. On our principles, that case has not been made.
Verdict
Watch. A knowable, cost-advantaged retailer with genuinely growing high-margin layers — but at these prices the margin of safety is not thin; it is absent, and management’s own buybacks were made well above our estimate of intrinsic value.
What would change it: confirmation, from the Q2 earnings call transcript or 10-Q, that excluding pharmacy deflation comps actually missed consensus (the 80bps versus 125bps ambiguity flips the sign), paired with either a sustained collapse toward the buy range or credible evidence that the high-margin businesses justify platform-level economics.
This analysis is AI-generated, educational, and not investment advice. Figures may contain errors or be delayed. Disclaimer