AXP

American Express

A closed-loop payments brand that underwrites the spender and bills the merchant — premium economics.

WideDecadeFullClassicBerkshire overlap

Longhold

Earnings

Moat

Hold

Ten-year thesis

Amex is not Visa. It is a brand plus a credit book plus a closed network, aimed at high-spend customers who want status and service. Credit risk is real (unlike Visa) and is the reason the earnings score is not a 95. A ten-year hold is a bet that affluent spend keeps growing and that the membership model survives 'buy now, pay later' and fintech cards. Buffett’s long ownership is a hint, not a substitute for underwriting the cycle.

The moat

Brand with affluent customers, closed-loop data advantage, and merchant willingness to pay for that customer. Switching costs via points and status. Credit cycle is the crack in the fortress.

Switching costs74

Pain, risk, or retraining required to leave.

Network effects70

The product gets better as more people use it.

Intangibles & IP78

Patents, data, licenses, process knowledge, regulation.

Scale advantage80

Fixed costs spread over a volume rivals cannot match.

Cost advantage68

Structural ability to be the low-cost producer.

Brand94

Pricing power from trust, habit, or identity.

Tech overlay

Tech is a feature (app, fraud models), not the moat. The moat is brand and closed-loop data. A tech-fluent owner can still like it without pretending it is a software company.

What can break it

Consumer credit downturn
Premium card competition (Chase, fintechs)
Merchant fee pressure

Earnings path

USD billions, illustrative 2016–2025 series

Owner metrics

P/E

18.0×

P/FCF

16.0×

ROIC

22%

ROE

32%

FCF conversion

75%

Debt / equity

1.70

Gross margin

55%

Op. margin

22%

10y rev CAGR

10y EPS CAGR

10%

Owner earnings

$8.5B

Mkt cap

$200B

Owner earnings yield 4.3%. Yield is judged against quality, not against a single hurdle — a fortress can be fairly priced at a lower yield than a fragile name.

PaymentsBrandBerkshire