Listen to the deep-dive discussion – Natural Versus Mechanical Corporate Growth (17:16 min)
“There is a crack in everything. That’s how the light gets in.”
— Leonard Cohen
I. The Spectrum Without a Name
The framework has defined two extremes.
On one end, the Freesurfer. Growth costs nothing. The wave arrives without capital expenditure, without human decision, entirely inside the moat. The toll and the wave fuse multiplicatively. The cost of B is zero. This is the Natural B.
On the other end, the mechanical compounder. Growth must be purchased. Every dollar of expansion requires a dollar of capital. Railroads spend billions to deepen their networks. GE invests $15 billion per engine program. The B is expensive, but when it stays inside the moat, it compounds reliably. This is the Mechanical B.
But between these two poles, there is a vast territory that the framework had not yet named. Most of the world’s greatest businesses live here. They are not Freesurfers—their growth is not entirely free. They are not purely mechanical—part of their growth arrives without effort. They are something else. Something imperfect.
This post names that territory.
———
II. The Dual Duration
Every business has two forces operating inside it.
The A—the toll. Cash now. Short duration. What the business collects today. Visa’s 0.2% on every transaction. S&P Global’s rating fee on every bond issuance. Coca-Cola’s shelf revenue this quarter. The A is the present tense of a business—the cash it generates by existing.
The B—the wave. Cash later. Long duration. The structural growth that arrives over time. The digitalization of payments for Visa. The expansion of global debt for Moody’s. The B is the future tense—the growth that compounds the A into something larger.
In The Infinite Investor, the original framework classified compounders into three categories. The A businesses—Coca-Cola, consumer staples—where the toll was dominant and the growth was modest. The B businesses—Visa, MSCI—where the wave was the engine. And the Dual businesses—Google, Berkshire Hathaway—where the A generated the cash that financed the B. Three fixed categories.
Then the Freesurfer emerged, and the framework evolved. The discovery was that the A and B are not separate categories of businesses. They are two forces present in every business, in varying proportions and varying degrees of quality. The question shifted from “is this an A or a B?” to “how good is this A, and how good is this B?”
The Freesurfer answered the question at the extreme: when the A and B fuse perfectly, when the toll is the wave and the wave is the toll, the business compounds geometrically. The short duration and the long duration become one. This is the A paramount.
But what about the businesses where the two forces don’t fuse? Where the A is strong and the B exists, but the B is partly free and partly costly? Where the short duration generates cash and the long duration requires some of that cash to be spent?
That is the Dual of the original framework—matured into something more precise. The Dual was an intuition. The Imperfect B is its full expression.
———
III. The Imperfect B
An Imperfect B is structurally divided. Part of the growth is natural—it arrives without capital expenditure, driven by secular forces that no one controls. Part is mechanical—it must be purchased, built, invested in.
The quality of an Imperfect B depends on two things.
First: the ratio. How much of the B is natural versus mechanical? A business where 75% of growth arrives for free and 25% requires capex is fundamentally different from one where the proportions are reversed. The higher the natural component, the closer the business is to a Freesurfer. The higher the mechanical component, the closer it is to a traditional compounder.
Second: the anchoring. Is the mechanical portion inside the moat? This is Bruce Greenwald’s insight applied at the sub-business level. If the mechanical B falls outside the franchise—on a level playing field, in a race—the cost is almost irrelevant. Capital spent outside the moat earns no excess return. But if the mechanical B is anchored inside the moat—if the data, the infrastructure, the relationships that fund the mechanical growth are themselves protected by the franchise—then the crack has light in it.
The Imperfect B is not a flaw. It is a spectrum. And the spectrum has a top and a bottom.
———
IV. The Archetype: Alphabet
If there is one company that defines the Imperfect B, it is Alphabet.
Start with the A—the toll. Google Search has over 90% of the global search market. “Google” is a verb. This is not a switching cost moat like Microsoft 365, where customers stay because migrating is painful. This is a behavioral moat—people stay because the alternative doesn’t exist at the same scale. No one signed a contract with Google. No one is locked in. The moat is closer to Visa’s—infrastructural, embedded in human behavior—than to any enterprise software franchise.
YouTube is a second A, independent of the first. A network of creators and viewers with effects that reinforce themselves. No contract. No lock-in. But no alternative either.
Together, Search and YouTube represent roughly 75% of Alphabet’s revenue. These two tolls are among the deepest and most durable in technology.
Now the B.
The natural B: Search and YouTube grow with the digitalization of global advertising—a secular wave that arrives without capex. More internet usage means more searches, more video consumption, more ad revenue. Each new website, each new creator, each new user deepens the flywheel without Google spending a dollar to acquire them. This B is not perfectly free—Google must maintain its infrastructure—but the marginal cost of growth approaches zero. Every additional search query costs nearly nothing and generates revenue inside the moat.
The mechanical B: Google Cloud requires massive capital expenditure on a level playing field shared with AWS and Azure. No single player has a structural moat in cloud computing. AI requires race-level investment with uncertain returns—Google competes with OpenAI, Microsoft, Meta, and a growing ecosystem of well-funded players.
This is the crack. Cloud and AI are the mechanical portion of Alphabet’s B.
But here is what makes Alphabet’s crack different from Microsoft’s: the mechanical B is partially anchored in the moat. Google has the data. Every search query, every YouTube interaction, every Maps request, every Gmail exchange, every Android device—these generate a reservoir of information that neither Amazon nor Microsoft possess at the same scale. When Google invests in AI, it feeds models with data that its franchise generates. The mechanical B is not floating on a level playing field. It is tethered to the A.
And the A pays for it. Search and YouTube generate so much cash that Cloud and AI capex can be funded entirely from operations—without borrowing, without dilution, without diverting resources from the core franchise. The natural A finances the mechanical B. This is the Dual of the original framework, fully realized.
The ratio is favorable—75% natural, 25% mechanical. The anchoring is partial but real. The self-financing condition holds. This is not purity. A Freesurfer would not need to spend on Cloud or AI at all. But it is high-quality imperfection.
There is a crack in everything. But in Alphabet’s case, the light gets in.
———
V. The Illusion of the Free B
Before going further, a distinction that prevents a common confusion.
Warren Buffett, speaking about Apple—the most valuable position in Berkshire’s history—said: “We did not do a damn thing.”
From Buffett’s chair, Apple’s B appears free. He bought the stock and watched it compound. No effort. No capital deployment. No decision. The growth arrived as if by magic. This looks like a Freesurfer.
It is not.
The framework looks at the structure of the business, not the experience of the investor. At the business level, Apple’s B is 100% mechanical. Tim Cook must execute every quarter. Design new products. Manage the most complex supply chain on earth. Innovate on hardware, software, and services simultaneously. If Cook stops executing, the growth stops. Apple’s B depends entirely on the racer.
What Buffett experiences as a “free ride” is actually the output of Cook’s relentless mechanical effort. The pacer surfs the work of the racer. It resembles a Freesurfer. It is not one.
The test is simple: if the CEO disappeared tomorrow, would the growth continue?
For Visa—yes. The transactions flow without a CEO. For Google Search—yes. People would still google. For Apple—no. Without the racer, the machine stops.
Apple is a brilliantly executed Mechanical B that creates the illusion of a free B for the passive investor. The Imperfect B is something structurally different—the growth is literally divided between a natural component and a mechanical component at the business level itself. Google Search grows by itself. Google Cloud does not. The division is architectural, not perceptual.
———
VI. The Quality Test
The Imperfect B is a spectrum. Google sits near the top. Other businesses sit lower. To classify any Hybrid, ask three questions in order.
Question one: What is the ratio of natural to mechanical B?
The higher the natural component, the closer the business is to a Freesurfer. Alphabet at roughly 75/25 is near the top. A business where mechanical growth dominates—say 30% natural and 70% mechanical—sits closer to a traditional compounder. The ratio is the first filter.
Question two: Is the mechanical B inside the moat?
This is the Greenwald test applied to the mechanical portion specifically. Google’s Cloud and AI spending is partially anchored by its data moat. The mechanical B is tethered to the franchise. Microsoft’s Azure and AI spending lands on a level playing field—three players, no structural moat. The same dollar of mechanical B, but one is inside the franchise and the other is outside. This distinction, as the previous post demonstrated, is what led Chris Hohn to cut 84% of his Microsoft position while increasing Alphabet.
Question three: Does the A generate enough cash to self-finance the mechanical B?
This is the test of sustainability—and the mature form of what The Infinite Investor originally called the Dual. The A pays for the B. Google’s Search and YouTube generate so much cash that Cloud and AI capex can be funded entirely from operations. The mechanical B is financed by the natural A, without borrowing, without dilution.
Microsoft’s M365 generates enormous cash—but $190 billion in annual capex strains even that franchise. The scale of the mechanical B is beginning to exceed what the A can comfortably finance. The self-financing condition—the Dual—is breaking.
A high-quality Imperfect B passes all three: favorable ratio, mechanical B inside the moat, self-financed by the A. A low-quality Imperfect B fails one or more. The test is sequential—if the ratio is unfavorable, the other questions barely matter. If the mechanical B is outside the moat, the cost is irrelevant.
———
VII. The Gradient Moves
A critical property of the Imperfect B: it is not fixed.
A Freesurfer’s position on the gradient is stable. Visa’s cost of B is zero today and will be zero in ten years. The ratio doesn’t shift because there is no mechanical component to grow. The purity is structural.
An Imperfect B can drift. The ratio of natural to mechanical can change as a company evolves its capital allocation. The anchoring can weaken if the moat erodes. The self-financing condition can break if the mechanical B scales faster than the A.
Microsoft is the case study. In 2019—cloud-first, M365-driven, capital-light—Microsoft’s Imperfect B was high-quality. The mechanical component was growing but modest relative to the natural cash flows of the franchise. By 2026, the mechanical B has exploded to $190 billion annually, much of it outside the moat. The ratio shifted. The anchoring weakened. The self-financing condition is strained. Microsoft descended the gradient—not because the A changed, but because the mechanical B grew faster than the moat could contain it.
The reverse can also happen. A business that reduces its mechanical B—through spinoffs, divestitures, or strategic focus—can ascend. S&P Global spinning off Mobility removed a mechanical B and left behind four natural B divisions. The ratio improved. The purity increased. The gradient moved up.
This is why Hohn re-evaluates every quarter. The Freesurfer is permanent. The Imperfect B is a snapshot that must be re-measured at every capital allocation cycle.
———
VIII. The Complete Map
The full spectrum, from top to bottom.
Natural B (Freesurfer). Growth costs nothing. 100% inside the moat. The A and B have fused—the short duration and the long duration are one. Visa, S&P Global, Moody’s, MSCI. The gradient does not move.
High-quality Imperfect B. Natural component dominates. Mechanical component is anchored in the moat. The A self-finances the B—the Dual realized. Alphabet is the archetype. The crack has light in it.
Mid-quality Imperfect B. Natural component significant but mechanical component growing. Anchoring partial or uncertain. The ratio is in flux. A business ascending or descending the gradient—the direction matters as much as the position.
Low-quality Imperfect B. Mechanical component dominates. Growth increasingly outside the moat. The Dual is breaking—the A can no longer comfortably finance the B. Microsoft 2026 at $190 billion of annual capex, majority on a level playing field. Descending.
Pure Mechanical B, inside the moat. Growth is expensive but 100% captured by the franchise. Railroads, GE Aerospace. The B is costly and the moat is deep. No crack—but no free ride either.
Mechanical B, outside the moat. Growth is expensive and falls on a level playing field. No franchise to protect the return. The worst position on the gradient. Capital competing on equal terms.
Every publicly traded company sits somewhere on this map. The Freesurfer is rare—five names in the entire market. The pure mechanical monopoly is uncommon. The vast majority of investable businesses are Imperfect—some natural, some mechanical, in varying proportions, with varying degrees of anchoring.
Knowing where a business sits on the spectrum—and which direction it’s moving—is knowing most of what matters.
———
IX. The Crack and the Light
The framework began with a simple taxonomy: A businesses, B businesses, and Duals. It evolved through the discovery of the Freesurfer—the rare case where the A and B fuse perfectly, where short duration and long duration become one, where the cost of growth is zero.
The Imperfect B completes the map. Between the purity of the Freesurfer and the weight of the mechanical compounder, there is a spectrum where quality is measured by three tests: the ratio of natural to mechanical growth, the anchoring of the mechanical component inside the moat, and the ability of the A to self-finance the B.
The Freesurfer is the ideal. But most investors will hold businesses that are less than perfect. The Imperfect B is the territory where the best investment decisions after the Freesurfer are made. The question is never whether a business has a crack. Every business below the Freesurfer does. The question is whether the crack lets the light in.
Alphabet’s crack lets the light in. The natural A is deep. The natural B is dominant. The mechanical B is partially anchored. The Dual holds—the A finances the B.
Microsoft’s crack is widening. The mechanical B is growing faster than the moat. The Dual is straining. The light is dimming.
The gradient within the gradient is not static. It is a living system. The Freesurfer sits above it all, permanent and pure. The Imperfect B lives inside it—moving, shifting, sometimes ascending, sometimes falling.
There is a crack in everything. The investor’s job is to find the cracks where the light gets in.
———
Watch the Video
Related posts on averagingup.com:
Hohn vs Ackman: Why $190 Billion Changes Everything · The Purification Trade · The Freesurfer · Building the Ark vs. Climbing the Mountain · Growth That Pays You · The Compression Harvest · The Permanent Wave
Based on the framework from The Infinite Investor, available at averagingup.com.
Averaging Up · May 2026 · averagingup.com