Lesson 5 of 7

The five sources of a moat

Step 3 of 6

Worked example

Three companies, three sources. Each claim is from the company's own 10-K, checked on 20 Nov 2026.

Visa: network effect

Visa runs the network connecting card holders, their banks, shops and the shops' banks. Its 10-K says "We do not issue cards, extend credit or set rates and fees for account holders of Visa card products."

The scale, for fiscal 2026 (year to 30 Sep 2026), in the 10-K's words: "total payments and cash volume was $18 trillion", with "over 5 billion payment credentials… available to be used at more than 400 million seller locations."

The test. Shops accept Visa because billions of people carry it; people carry it because almost every shop accepts it. A new network starts with neither.

What it shows up as. Visa's FY2026 net revenue was $45,903M and operating income $28,110M.

operating margin = operating income ÷ revenue
                 = 28,110 ÷ 45,903
                 ≈ 61.2%

About 61 cents of each revenue dollar was left after running the business: consistent with a strong moat, though a margin alone can't prove which source causes it.

Nvidia: switching costs

Nvidia sells GPUs (graphics processing units: chips that do many small calculations at once, which is what AI work needs). Yet its 10-K leads with software: "the foundational NVIDIA CUDA development platform that runs on all NVIDIA GPUs", introduced in 2006, with "support for 6,000 applications". CUDA is Nvidia's own toolkit for writing programs that run on its chips.

The test. A team with years of code written for CUDA would have to rewrite and retest it to move to a rival's chips. That cost is the switching cost. Nvidia's version: "The large and growing number of developers and installed base across our platforms strengthens our ecosystem."

What would weaken it. The 10-K lists "software support" and "conformity to industry standard APIs" among the main ways chip makers compete. An API is an agreed way for programs to talk to a chip, like a standard plug. If rivals' chips could run the same programs through standard plugs, with little rewriting, customers could move cheaply and the switching cost would shrink. (The 10-K gives no developer count, so don't quote one.)

Costco: cost advantage

Costco's 10-K: low prices on "a limited selection" of products produce "high sales volumes and rapid inventory turnover". With bulk buying and "no-frills, self-service warehouse facilities", this lets it "operate profitably at significantly lower gross margins… than most other retailers."

The numbers you already have show it working. Costco keeps about 11 cents of each sales dollar after paying for the goods (lesson 01). Its gross margin, fees included, was 12.8% against 33.2% for general retailers (Module 6 lesson 02), yet it earned a 22.9% return on invested capital (Module 6 lesson 03), because it turns its assets over 3.65 times a year (Module 6 lesson 04). A rival charging Costco's prices without that volume would have the thin margin without the turnover.