Listen to the deep-dive discussion – The Only Five Perfect Monopolies (17:19 min)
“The most important thing is high barriers to entry. If you find a company that is going to be a good company long-term, then you should hold on to it because there’s a persistency of these barriers to entry.” — Chris Hohn, interview with Nicolai Tangen, May 2025
* * *
I. Beyond Depth
Every investor asks: how deep is the moat? Morningstar rates them narrow or wide. Buffett demands them before he writes a cheque. Dorsey catalogued them — network effects, switching costs, intangible assets, cost advantages. The moat is the most discussed concept in investing.
But almost no one asks the second question: what sustains the moat?
A moat can be deep and still be built on sand. If the force that created the moat is a human decision — a regulation, a mandate, a political directive — then another human decision can reverse it. In one afternoon. With one tweet. And thirty years of monopoly dissolve.
This is what happened to FICO in 2025. And the lesson it teaches reaches far beyond one company. It reveals a hierarchy within the concept of the moat itself.
* * *
II. FICO Was a Freesurfer
In the Averaging Up framework, a Freesurfer is defined by the fusion of A and B. The A is the toll — cash now, short duration, what the business collects today. The B is the wave — cash later, long duration, structural growth that arrives without human decision, carried by external forces. When the B is natural and gratuitous, and the toll captures 100% of the wave, the two fuse. A × B. The sign is multiplicative. The business compounds geometrically. This is the A paramount.
FICO was a textbook case. The toll: every credit decision in America required a FICO score. Mortgage, auto loan, credit card, personal loan — the score was the gatekeeper. The wave: the financialization of the world. More credit, more consumers, more countries adopting credit scoring, more decisions automated by AI. FICO did not build this wave. The world financialized for its own reasons. The B was natural, structural, gratuitous. Uninvited, unpaid for, unannounced.
And the moat appeared absolute. Fannie Mae and Freddie Mac — the two government-sponsored entities that purchase the majority of American mortgages — accepted only FICO scores. Not by market choice. By regulatory mandate. The moat was inscribed in the rules of the system.
Operating margins of 55%. ROIC stratospheric. Capital-light. PE of 80. The market priced FICO as if the fusion were permanent.
* * *
III. The Breach
In mid-2025, FHFA director Bill Pulte authorized Fannie Mae and Freddie Mac to accept VantageScore 4.0 as an alternative to FICO for government-backed mortgages. One regulatory decision. One afternoon.
VantageScore is jointly owned by Equifax, TransUnion, and Experian — the three major credit bureaus. They had created it in 2006 as a competitor to FICO, but it had been locked out of the mortgage market by the regulatory mandate. The moment the mandate fell, the war began.
TransUnion slashed VantageScore pricing to $0.99 per score. Equifax to $1. Experian to $0.99. All three offered VantageScore for free to any lender already purchasing a FICO score. Against FICO’s $10. JPMorgan and Bank of America tested both models on millions of consumer mortgages and found VantageScore 4.0 “better” at predicting credit risk.
FICO’s stock lost 50% from its 52-week high. On March 12, 2026, it plunged another 9% in a single session after the coordinated pricing offensive.
The exclusive toll became a contested toll. And the Freesurfer began to retrograde.
* * *
IV. The Mechanism of Retrogradation
The natural B at the industry level did not disappear. The financiarization of the world continues. Credit expands. Scores are needed. The wave is still there, still structural, still gratuitous. No one builds it. No one controls it. It arrives uninvited.
But the natural B at the firm level changed nature.
When FICO was alone on the reef, 100% of the wave passed through its toll. The industry’s natural B was FICO’s natural B. The two were indistinguishable. That indistinguishability was the fusion. That fusion was the A paramount.
The moment VantageScore appeared at $0.99, the wave split. FICO now had to fight for its share of the wave — cut prices, launch its Mortgage Direct licensing program, buy back $1.5 billion of its own stock to reassure the market, invest in FICO 10T to prove its model’s superiority. Every dollar spent retaining the wave is mechanical B. Human decisions. Execution risk. Capital deployed.
The seam between A and B reappeared. The fusion dissolved. A × B reverted to A + B. The sign changed from multiplicative to additive.
THE PARADOX
The B remains natural for the industry but becomes mechanical for every participant. FICO must fight. VantageScore must fight. Equifax gives scores away for free. TransUnion at $0.99. Experian at $0.99. Each participant converts the industry’s natural B into firm-level mechanical B. The wave is free. Surfing it no longer is.
The Freesurfer existed because one actor captured 100% of a natural B. The moment two actors share the same B, both become mechanical. The exclusivity of the toll on the wave is the necessary condition of the fusion. Without exclusivity, there is no fusion. Without fusion, there is no Freesurfer.
THE INDISSOCIABILITY
The critical insight: it is not the B that FICO lost. The wave is still there. It is the exclusivity of the A’s claim on the wave that eroded. And a breach in the A, by allowing competition to share the B, converts the B from natural to mechanical in return. The A and the B are indissociable. You cannot lose one without losing both.
FICO did not die. It retrograded on the gradient. From A × B natural to A + B where the wave is still natural but the captation is contested. It still has 90% market share. It still generates strong cash flow. Its Q1 2026 earnings were excellent — revenue up 16%, scores revenue up 36%. But it is now dependent on human decisions and execution to maintain its access to a wave that was once exclusively its own. Pricing strategy, product innovation, direct licensing. The A is in contraction. The moat costs more to defend each quarter.
* * *
V. The Hierarchy of Moats
FICO reveals a hierarchy within the concept of the moat. Not all moats are equal. The discriminant is not how deep the moat is. It is what sustains it.
REGULATORY MOAT
Created by human decision — a law, a directive, a mandate. What a human decision creates, another human decision can destroy. A director of the FHFA, a tweet by Bill Pulte, a vote in Congress. The moat is as strong as the political will that sustains it. And political will is the most volatile variable in finance. The reef is made of sand. One tide and it dissolves.
MECHANICAL MOAT
Sustained by continuous human decisions within the company — $4 billion in annual advertising (Coca-Cola), $15 billion in R&D per cycle (ASML), acquisitions (Couche-Tard), capital allocation (Berkshire). The moat is real and often very deep. But the day the human stops deciding — the day the CEO misallocates, the day the R&D falters, the day the acquisition fails — the moat erodes. It is deep but on life support. Human life support.
AUTO-REINFORCING MOAT
Emerged from the structure of the system itself. Network effects, infrastructure lock-in, standard adoption, switching costs embedded in billions of integrations. Born mechanical — every moat begins with human decisions accumulated over decades. But the accumulation passed a tipping point. The system now perpetuates itself without human decision. Visa’s network — billions of cards, millions of merchants, decades of integration — reinforces itself with every transaction. No one decreed that Visa would process 60% of global card transactions. No one can decree that it stop. To destroy this moat, you would have to destroy the system itself. The reef is made of coral — living, growing, impossible to move.
This is why Visa is superior to FICO in the gradient. Not because the B is more natural — both waves are equally structural. Not because the toll is higher — both collect a percentage on every transaction in their respective domains. But because the moat is of a different nature. Visa’s moat is coral. FICO’s was sand. The same wave hit both. One stood. One dissolved.
* * *
VI. Dual Duration
The Infinite Investor decomposes every business into two forces that operate on different time horizons.
A = short duration. Cash now. The toll. What the business collects today. The A is always short by definition. What varies is the moat that protects it — which determines whether the A will repeat tomorrow. The moat is a property of the A, not a separate variable. It qualifies the A’s repeatability.
B = long duration. Cash later. Growth. What the business will collect tomorrow.
Some exceptional companies contain both durations internally. They generate massive current cash flows (A) and invest heavily in long-duration growth (B). Berkshire Hathaway, Alphabet, Amazon — each does internally what the investor tries to do at the portfolio level. The A finances the B. The short duration funds the long duration. These are Dual Duration companies.
But not all Dual Duration companies are equal. The difference lies in the nature of the B — and this distinction is the key to everything that follows.
A mechanical B is growth built by human decisions. Every dollar of growth requires capital, execution, and continuous human judgment. R&D spending, acquisitions, new store buildouts, capital allocation. The growth depends on someone deciding correctly, repeatedly, over decades. If the decision-maker falters — a bad acquisition, a misallocated billion, a CEO transition — the B falters with it. ASML spends $15 billion per cycle to maintain its technological lead. Couche-Tard acquires networks of convenience stores one by one. Berkshire deploys capital through the judgment of a single allocator. These are excellent businesses. But their growth is human-dependent. Flawless and continuous execution is the price of the mechanical B.
A natural B is growth carried by external structural forces. No one inside the business decides it. No one builds it. It arrives because the world moves in a direction that funnels volume through the toll — uninvited, unpaid for, unannounced. The digitalization of payments sends more transactions through Visa’s rails. The shift to passive investing sends more assets under MSCI’s benchmarks. The expansion of global debt sends more issuance through the rating agencies. These forces are structural, secular, and independent of any decision by the business itself. The cost of the natural B is zero — not because someone decided to make it free, but because it is free by nature.
The sign between A and B — additive (+) when growth costs something, multiplicative (×) when it is gratuitous — determines whether the business compounds arithmetically or geometrically.
And the summit: when the moat is auto-reinforcing and the B is natural, the double naturality is achieved. The moat costs nothing to maintain. The growth costs nothing to produce. The entire business functions without human decision — neither to protect what it has nor to grow what it will become. This is the A paramount.
* * *
VII. The Complete Taxonomy
Eight stable categories. From the basement to the summit.
| Category | Moat (repeatability of A) | Nature of B | Sign | Examples |
| Sub-A (no moat) | None. The A stops if the human stops. | Absent. | — | The café with two locations. The solo law practice. The independent convenience store. |
| A well operated, no moat | None structural. Good execution without protection. | Absent or negligible. | — | Biosyent. Canadian Tire. A well-run regional chain. |
| Stalwart (A, mechanical moat) | Mechanical. Repeatability conditional on continuous execution. | Compounding at steady rates. Returns cash to shareholders. Short duration. | + | Coca-Cola. Walmart. Johnson & Johnson. |
| B alone | No A. No toll. | Pure narrative. Wave without reef. | — | Pre-revenue startup. Crypto. Meme stock. |
| A + B mechanical | Mechanical. Continuous execution. | Mechanical. Every $ of growth costs capital. | + | ASML. Couche-Tard. Berkshire. |
| A + B mechanical (auto-reinforcing moat) | Auto-reinforcing. Network lives on its own. | Mechanical. Growth costly despite excellent moat. | + | Amazon (network + warehouses). |
| A + B partially natural | Mechanical. | Partially natural. Two sub-cases: cyclical tailwind (not secular), or secular wave with contested captation. | + | Dollarama (cyclical trade-down). FICO post-2025 (secular wave, contested captation). |
| A × B natural — A Paramount | Auto-reinforcing. Unconditional. Coral. | Natural, gratuitous, uninvited. The wave. | × | Visa. Mastercard. S&P Global. Moody’s. MSCI. Five on Earth. |
Note on retrogradation: a Freesurfer whose moat erodes — particularly a regulatory moat reversed by human decision — descends the gradient. FICO passed from category 8 (A × B natural) to category 7 (A + B naturally present but captation contested) when the FHFA opened the door to VantageScore. The wave is still there. The access to the wave now costs. The taxonomy describes stable categories. The mechanism of descent between them is the retrogradation described in Section IV.
* * *
VIII. Why Five on Earth
The A paramount requires both naturalities simultaneously. The moat must be auto-reinforcing — the system perpetuates itself. The B must be natural and gratuitous — external structural forces carry the growth for free. This double naturality is extraordinarily rare.
Many businesses have an auto-reinforcing moat. Amazon’s marketplace network effect is as autonomous as Visa’s payment network. But Amazon’s B is mechanical — every warehouse, every delivery route, every data center costs capital. The moat is coral. The B is not free. Half the equation.
Many businesses have a partially natural B. Dollarama benefits from trade-down when inflation rises — customers arrive without invitation. But it is cyclical, not secular. And Dollarama’s moat is mechanical — each new store is a human decision. A tailwind is not a wave. Half the equation.
Many businesses had a complete equation that proved fragile. FICO had the fusion — exclusive toll on a structural wave. But the moat was regulatory. Sand. The fusion was conditional on a political decision persisting. It did not.
The Big 5 — Visa, Mastercard, S&P Global, Moody’s, MSCI — are the only businesses where the moat is auto-reinforcing AND the B is natural. The network lives on its own. The wave arrives on its own. The toll collects on its own. No human decision needed. Neither to protect nor to grow.
The coral reef in an ocean current. That is why there are five on Earth.
* * *
The moat that matters is not the one you can see. It is the one that no one can reverse.
* * *
This post is part of the framework described in The Infinite Investor, available at averagingup.com. For the complete architecture — including the Freesurfer taxonomy, the cost of B, the Infinite Option, the Bucket system, and position sizing — see the book.
Related posts: “The Freesurfer” — “The Big One” — “What Chris Hohn’s Portfolio Reveals” — “Position Sizing: The Final Act of Conviction” — “The Double Swell Freesurfer”