How a moat dies

A moat rarely fails suddenly. It fails in one of four ways, and three of them are visible in disclosures years before they show up in a margin.

Ending one: exclusivity runs out

A patent has an expiry date, published years in advance, and it is the only moat failure that arrives on a schedule. The revenue decline that follows is not gradual.

The Bristol Myers Squibb figures above are what that looks like in a filing. Revlimid revenue of 2,951 million dollars in 2025 against 5,773 million in 2024. Sprycel at 493 million against 1,286 million. Abraxane at 368 million against 875 million. Total revenues across the company were essentially flat at 48,194 million against 48,300 million, because other products grew into the gap.

The aspartame case makes the same point from the other side: after the European patent expired in 1987 and a competitor entered, the price of aspartame fell 60 percent, and the entrant lost money doing it.

Ending two: substitution

The 1991 Berkshire letter documents one in progress. It observes that the economic strength of media businesses was eroding as retailing patterns changed and advertising and entertainment choices proliferated, and that newspaper, television and magazine properties had begun to resemble businesses more than franchises in their economic behaviour.

Note what changed and what did not. The newspapers still printed, the advertisers still advertised, and revenue in the year was not the evidence. The evidence was that customers had acquired alternatives, which meant the second of the three franchise conditions, that customers believe there is no close substitute, had stopped holding.

And the letter prices it: a property valued as a growing perpetuity at 25 times after-tax earnings becomes, on a bob-around assumption, worth 10 times. The valuation consequence of a substitution arrives long before the revenue consequence does.

Ending three: the scale stops mattering

A scale advantage rests on a market being the unit of competition. When that stops, the share that created the advantage stops producing anything.

This ending is the hardest to see because nothing in the company's own numbers changes at first. What changes is that a competitor can now reach the customer without paying the cost the incumbent's share was spreading.

Ending four: the owner's own decisions

An acquisition into an adjacent business with no barrier. A price increase that teaches customers to look elsewhere. Cost cutting into the thing the customer was paying for. Extending the brand until it stops meaning anything specific.

This one leaves the clearest trail, because every part of it is disclosed: acquisitions in the business combination note, pricing in the revenue discussion, and cost cutting in the expense lines. It is also the only ending where the company chose it.

Check yourself