From Toll Roads to Cloud Platforms — The Business Model That Never Dies.
Yesterday we wrote about Microsoft's toll road strategy — how the company is building five roads that funnel every AI model through its billing, security, and data layer before reaching enterprise customers. Today, we ask a different question: is this strategy actually new?
It is not. It is one of the oldest business models in economic history. The assets change. The logic does not. Whoever controls the passage that others must travel through collects money without needing to own what passes through.
Three historical examples make this clear. Each one operates at a different scale — sea, canal, and smartphone. But the structure is identical every time.
The first is the most ancient version: the Oresund Strait.
For over 400 years, Denmark collected tolls from every ship entering or leaving the Baltic Sea. The strait between Denmark and Sweden was narrow enough that Denmark could enforce payment. UNESCO's Sound Toll Registers — one of the most detailed commercial archives in European history — document this flow in extraordinary detail: ship names, cargo types, origin ports, fees paid, spanning from the 15th century onward.
Denmark did not build the best ships. Denmark did not own the cargo. Denmark owned the bottleneck. That was enough. For centuries, it was one of the Danish crown's largest revenue sources. The lesson is not subtle: you do not need to manufacture what moves through your territory. You need to make your territory unavoidable.
The second example is the modern industrial version: the Suez and Panama Canals.
According to Britannica, the Suez Canal became the shortest sea route between Europe and Asia and one of the most heavily used shipping lanes in the world. The Panama Canal achieved the same for Atlantic-Pacific transit. Both are pure toll businesses. The canal operator does not own the containers. It does not manufacture the goods inside them. It sells a shortcut — and because geography makes the alternative so much worse, the shortcut commands a premium every single day.
UNCTAD confirms that maritime transport still carries over 80% of global merchandise trade by volume. This is not a relic of the past. It is the present. The canals that sit on those routes are still printing money from the same structural logic they were built on.
Now, the question people never ask: what is the modern version of a canal? What is today's unavoidable shortcut that commands a toll?
The third example answers that: Apple's App Store.
Apple's official developer documentation confirms a standard 30% commission on paid apps and in-app purchases. The App Store Small Business Program reduces that to 15% for developers earning under $1 million annually. Apple did not build every app. It did not need to. It built the distribution layer — discovery, payment processing, trust certification, policy enforcement, tax collection — and wrapped it into a single gateway that 1.5 billion device owners pass through by default.
The structure is the same as the Oresund Strait. The same as Suez. Control the passage. Let others build what travels through it. Collect a percentage of everything.
Now here is where intellectual honesty matters. Drawing a straight line from sailing ships to cloud computing and calling it a law of history is overreach.
Three reasons.
First, new distribution layers rarely replace old ones completely. They stack. UNCTAD's data proves this — maritime shipping still dominates global trade by volume even in the age of air freight and digital commerce. IATA notes that air cargo carries a small fraction of volume but a large share of value. The old roads do not die. New ones are built on top.
Second, the categories are not the same. Roads and canals are physical transport. The internet and wireless are communication infrastructure. Smartphones are consumer terminals. AI is a service and control layer that sits on top of all of them. Calling them all "toll roads" is a useful metaphor, not a precise equivalence.
Third, AI is not yet a fully independent distribution network. It is closer to an integrated control layer riding on existing infrastructure — cloud, fiber, data centers, power grids. It is less a new road and more a unified command center built above the roads that already exist.
So the honest framing is this: the toll road metaphor is directionally correct but structurally incomplete. The pattern — control the chokepoint, collect from the flow — repeats across centuries. But each era's chokepoint operates differently, and assuming a clean linear evolution overstates the case.
What does this mean for Microsoft specifically?
Microsoft is not trying to be the next Denmark or the next Apple — at least not in the consumer world, where it has already lost the mobile entry point to Apple and Google. What Microsoft is doing is building the enterprise version of this structure. Azure, Foundry, Fabric, Copilot, and the security stack are not consumer toll gates. They are enterprise toll gates — designed to become the default path that companies use to deploy AI into their operations.
If that path becomes the most natural, the most secure, and the least painful option for enterprises, then the same structural logic applies. The customer does not want to leave. The switching cost is not a fee — it is the pain of rebuilding identity, security, data pipelines, compliance, and workflow integrations from scratch.
History suggests that whoever builds that kind of path tends to collect rent for a very long time.
But history also suggests that no toll collector is permanent. Open standards can erode lock-in. Rival platforms can replicate the structure. Regulatory action can force interoperability. The moat is real, but it is not infinite.
Our view — same as always. This is not investment advice for any single stock. Microsoft's FY26 Q3 earnings arrive April 29. If you are watching this company, watch the enterprise AI deployment numbers, not the model benchmarks. The model race makes headlines. The distribution race makes revenue. Do not use leverage. Do not take loans. Cash earning 3.5% in a money-market fund is a position, not a surrender. The goal is not to predict the winner. The goal is to still be standing when the answer becomes clear.
Related Posts:
