◆ What the Market Isn't Pricing In

Costco is valued as a retailer that happens to be excellent. The accounts say it is two companies, and the market has never really priced the better one separately.

Half the operating income is a subscription. Membership fees were $5,323 million of $10,383 million in fiscal 2025 and $4,057 million of $7,884 million over the first thirty-six weeks of fiscal 2026¹. That half has a 92.2% renewal rate, no cost of goods, no inventory risk, no freight exposure, no shrink, is collected twelve months in advance, and is repriced upward roughly every six years without measurable attrition. If it were listed on its own it would be one of the highest-quality subscription businesses in the world and would not trade at anything like a retailer's multiple.

The other half is a genuinely ordinary retail business — $5,060 million of profit on $269,912 million of sales, a 1.87% margin, exposed to everything retail is exposed to. Consolidating them produces a company that looks like a retailer with an oddly high multiple, which is how most commentary treats it.

Three things follow that are visible in the filings and rarely discussed.

One is that the subscription is being underpriced on purpose. The fee has been raised roughly every five to six years — most recently from $60 to $65 and $120 to $130 on 1 September 2024² — which means $65 today buys less than $60 did in 2017. Costco has spent three decades letting the real price of its highest-margin product fall. That is not a weakness; it is an unexercised reserve of several hundred million dollars a year, refreshed continuously by inflation, and the 92.3% renewal rate after the last increase is the evidence that it exists.

Another is what the Sam's Club comparison actually proves. Walmart has run the identical format for forty-three years and in fiscal 2026 its American club segment's operating expenses of $10,639 million exceeded its gross profit of $10,556 million³ — the merchandise business loses money, and all of the segment's profit is dues. The variable is not merchandising skill or buying power, both of which Walmart has in abundance; it is sales density, where Costco does roughly $295 million per warehouse against $155 million⁴. That is a controlled experiment with one variable, run over four decades, and it is the strongest evidence available that this moat is structural.

And then the balance sheet, which almost nobody counts. Members supply $3,157 million of deferred fees, suppliers supply more than the entire inventory — payables of $22,363 million against stock of $19,418 million — and the company sits on $19,996 million of cash against $5,670 million of long-term debt⁵. Roughly $20 billion of the market capitalisation is cash, which means the operating business is being valued at nearer $386 billion, and the 46 times becomes about 44.

What the market is pricing correctly is the growth rate, which is high single digits, and the risk that a 3.0% net margin leaves very little room. What it is arguably not pricing is that one of the two businesses inside this company almost never has a bad year — and that Costco has been deliberately declining to charge full price for it since 1983.