The World Still Runs Through Chokepoints

There’s a particular kind of confidence that comes from complexity.

When supply chains span dozens of countries, when energy flows through layers of intermediaries, when financial systems connect thousands of counterparties — complexity starts to feel like protection. That something this elaborate, this globally distributed, couldn’t possibly fail in a single place.

That assumption has had a difficult few months.

What we’re watching isn’t a geopolitical shock disrupting a resilient system.
It’s a resilient-looking system revealing that it was never as distributed as it appeared. The routes, the capacity, the pricing mechanisms, the insurance structures — these didn’t break under unusual stress. They buckled under conditions the system was always vulnerable to.

The problem wasn’t the shock. The problem was the story we told ourselves about what the system could absorb.

How Concentration Happened

The concentration didn’t arrive through negligence. It arrived through optimisation.

Over three decades, global trade settled into its most efficient geometry. Energy infrastructure oriented around the lowest-cost extraction and the shortest viable route to market. Shipping networks consolidated around a small number of high-throughput corridors because volume density made those routes cheaper, faster, and better serviced than alternatives. Insurance capacity deepened in a handful of specialist markets because expertise and capital tend to pool where risk is most legible and most frequently priced.

Each decision made sense individually. By conventional measures, each was rational. The problem is that rational local optimisation often produces irrational system geometry — and the geometry that emerged was, in several critical places, extremely narrow.

Consider what that narrowness looks like in practice. A significant share of global liquefied natural gas moves through a small number of maritime corridors. Fertiliser supply chains — and therefore food production capacity for billions of people — run through a limited set of export origins and processing nodes. Marine insurance for large portions of global trade is written by a market with genuine depth in perhaps four or five locations. Shipping capacity, when disrupted, cannot be rerouted freely — the vessels, the port infrastructure, the crewing arrangements, the flag state certifications — these don’t flex quickly.

This is not a diversified system with redundancy.

It is a system that looks diverse at the surface and runs thin underneath.
The concentration was, in every case, the logical outcome of the incentive structure. Which is precisely why it was so difficult to see.

"Globalisation didn’t distribute risk. It disguised it."

Why Organisations Missed It

Risk frameworks are not neutral instruments. They encode a worldview.
The worldview embedded in most enterprise risk frameworks — built and refined across the 1990s and 2000s — assumed several things about the operating environment that have since quietly expired. It assumed that major trade routes would remain functionally open. It assumed that specialist insurance markets would remain willing to price and bear risk on normal terms. It assumed that geopolitical actors, whatever their differences, shared a basic interest in maintaining the conditions for global commerce.

These weren’t naïve assumptions at the time. They were reasonable extrapolations from a long period of relative stability. The mistake wasn’t making them. The mistake was not revisiting them.

What this produced, in practice, was a generation of risk assessments that modelled volatility within a stable system rather than questioning the stability of the system itself. Scenario planning asked: what happens if this route is disrupted temporarily? It rarely asked: what happens if the assumption that this route is available at all turns out to be wrong?

I’ve watched risk teams run sophisticated stress tests — detailed, well-resourced, technically credible — that nonetheless couldn’t surface the vulnerability because the concentration wasn’t inside the model. It was in the model’s foundations. The question being asked was the wrong one. And frameworks are very good at answering the question they were built to ask.

This is the structural failure underneath the current disruption. Not that organisations lacked information. Not that signals were absent. But that the architecture of their risk thinking was optimised for a world that no longer exists — and that the gap between the assumed world and the actual world had been growing, quietly, for years before anything broke.

What a Chokepoint Actually Is

The result was a system that looked stable on paper and fragile in reality. To see why, it helps to be precise about the structure of a chokepoint.

The word is used loosely. It’s worth defining it properly.

A chokepoint, as I’m using it in this series, is a node in a critical system where three conditions converge: high dependency, low substitutability, and long response lag. These variables compound. High dependency means that a large volume of something essential flows through the node. Low substitutability means there is no viable alternative routing, sourcing, or capacity that can be activated quickly. Long response lag means that even if substitutes exist in theory, the time required to deploy them is measured in months or years, not days or weeks.

What makes a chokepoint dangerous — rather than merely constrained — is when a fourth condition is present: uncertain political will. When the entities capable of resolving a chokepoint disruption have interests that may not align with rapid resolution, the response lag extends further and the damage compounds in ways that most scenario models don’t price.

Hormuz is the canonical example. But Hormuz is canonical precisely because it has been visible for decades. The more dangerous chokepoints are the ones that haven’t yet been named in a board risk register — the systems where dependency is high, alternatives are thin, and the assumption of availability has never been seriously tested.

Globalisation didn’t distribute risk. It disguised it. It took physical concentration — narrow straits, single-origin supply lines, thinly capitalised insurance markets — and wrapped it in the language of diversification. Complexity became a proxy for resilience. The more elaborate the system looked, the harder it became to see where it actually ran thin.

This series examines what’s running thin. Not the headlines — those will keep moving. But the structural conditions underneath them: the places where critical systems depend on assumptions that quietly expired, and where most organisations’ frameworks were never designed to look.

The first thing worth understanding is how the concentration happened. Not suddenly, and not by accident.

“The opposite of a resilient system isn’t a fragile one. It’s one that performs resilience until the moment it can’t.”

View all posts in this series →