What We Get Wrong About Cities

No one doubts that Paris generates more than $1 billion a year in transport transactions. The metro alone moves more than 1.5 billion riders annually. Add taxis, ride-hail, buses, and commuter rail and the number is multiples of that. The math is obvious because it is measured.

Abidjan is not Paris. But Abidjan has roughly 6 million people, more than 160 transport companies managing over 100,000 drivers, and $90 million in annual fleet renewal. In 2025, AI-optimized routing saved commuters in Abidjan more than 815,000 hours — the highest of any African city we operate in. The math is there. It is just not measured.

This is the central error in how capital gets deployed across emerging markets. People look at countries. They should be looking at cities.

The population trap

Venture capital in emerging markets is supposed to create wealth, not just fuel consumption. But when investors use national population as the yardstick, decisions get made in a way that perpetuates consumption as the value driver. Nigeria has 220 million people. Pakistan has 240 million. The numbers look enormous. Capital flows in sized to those numbers.

The productive populations in both countries mostly reside in their main cities, not in the vast rural areas that make up the bulk of the headcount. Lagos, Abuja, Karachi, Lahore. Once you reframe around productive urban cores, the scale difference between Nigeria and Côte d'Ivoire shrinks from roughly 5:1 to something closer to 2:1. Suddenly, the market that looked five times larger is only twice the size — and the one that looked too small to matter is very much in play.

The distinction changes how much capital you raise, where you deploy it, and how fast you need to grow. Jumia tried to build e-commerce for Nigeria — the country, 200 million people. Raised hundreds of millions, IPO'd, and ended up with a market cap lower than the total private capital raised. The country-level thesis consumed the capital.

What a city actually tells you

At the country level, you are blinded by averages. Cities force you to think in median terms because cities are ecosystems of proximity — both physical and relational. Geoffrey West's research on urban scaling found that when a city doubles in population, its economic output increases by roughly 115 percent — not 100.¹ The productivity gains from density are superlinear, and they are invisible at the country level.

I landed in Abidjan and the first thing that struck me was how different the parts of the city were from each other. Abobo is a sprawling, dense neighborhood with informal everything. Plateau feels almost European — clean infrastructure, commercial density, functioning services. That gap tells you where the money is going. People in Abobo are not aspiring to live in Paris. They are aspiring to live in Plateau. The economic pull runs within the city itself, and you can see it in how people move, spend, and build.

In many African cities, even the affluent neighborhoods have broken infrastructure — unreliable electricity, poor plumbing, roads that flood. You cannot read the direction of economic motion because the breaks cut across income levels. Abidjan is different. The gradient is visible. And a visible gradient means predictable demand.

How to read a city in 72 hours

Country-level analysts have the Ease of Doing Business Index and GDP per capita tables. These are useless for operational decisions. A city demands a different set of instruments.

The first variable is transport node mismatch. Find where the city's bus terminals and transit stops are, then map where the youngest 20 percent of the workforce actually lives. If high-density residential clusters are forming in dead zones without official transport, the city is about to experience either a forced infrastructure pivot or a localized commercial boom. Both are opportunities.

The second is the specialty-to-staple ratio. Count the specialized services — repair shops, niche gyms, specialty coffee — relative to staple goods like groceries and fuel. This ratio is a proxy for discretionary spending power that no national statistic will give you. A neighborhood with three specialty coffee shops and a coworking space is telling you something about the median income of a five-block radius that the World Bank cannot.

The third is time-to-clearance for the smallest unit of trade. National data tracks business registration. A city-level analyst looks at how long it takes to set up a street stall, open a bodega, or get a shared workspace permit. This is the real friction coefficient of the local economy. If a woman can set up a food stall in a morning, the city has low transactional entropy. If it takes three weeks and four permits, the city is choking its own commercial metabolism.

I wrote about the full framework in The First 72 Hours in a New City. But the point here is simpler: these variables are invisible from the country level. You have to be standing in the city to see them.

Transport is a commodity. That is the point.

The thing people find hardest to internalize is that transportation is a global commodity. Cars are manufactured and shipped around the world at roughly similar prices. Fuel costs roughly the same. The physics of moving a person from point A to point B does not change because you are in Abidjan instead of Amsterdam.

What changes is how transactions happen. In cities where most economic activity is relational — where commerce requires people and goods to physically move for a transaction to occur — transport stops being a convenience and starts functioning as the transactional layer of the economy. Reduce the cost of movement and you reduce the cost of every transaction in the city.

The spectrum runs from communal buses and walking at one end to private cars and taxis at the other. As wealth is created in a city, people move along that line. The bus rider becomes the shared-taxi rider. The shared-taxi rider becomes the private-ride user. This progression follows the money, not the aspiration. Each step up means faster movement, more transactions per day, higher individual productivity. The 815,000 hours saved in Abidjan last year went straight back into the economy as freed productive capacity.

The same commodity logic that makes this work in Paris makes it work in Abidjan, in Kinshasa, in Dakar. In Kinshasa, AI routing improved trip efficiency by 6 percent — one of the strongest gains globally, in one of the cities most people in tech could not locate on a map. The gains are largest where the existing system is most inefficient — where the country-level lens tells investors not to look.

The data spine

A transport platform that reaches critical mass in a city does not just move people. It generates a continuous data layer — trip patterns, demand density, time-of-day flows, payment behavior — that makes every subsequent business cheaper to build. This is why transport is the entry point, not the destination.

Connect that data layer to vehicle financing and you get what I described in The Financing Loop — a self-reinforcing system where transport data makes vehicles scoreable, which makes banks willing to lend, which puts more vehicles on the road, which generates more data. The loop compounds. Each city that reaches critical mass becomes a platform for financial services, logistics, and commerce that would have been impossible to build from scratch.

This is why the city-level thesis matters operationally, not just analytically. You do not build a data spine at the country level. You build it city by city, each one reaching the density where the data becomes valuable enough to underwrite the next layer. The businesses that print money in emerging markets are the ones that fill institutional vacuums this way — not by disrupting what exists, but by becoming the infrastructure that did not exist before.

The real number

There are more than 500 cities in Africa, Latin America, the Middle East, and Central Asia with populations above 1 million.² Every one of them can accommodate a $100 million offline-to-online business. Not a hypothetical $100 million. A transport platform that becomes the data spine for lending, insurance, payments, and logistics — each layer financializing the one below it.

The aggregate is large enough that I will not write it here because it sounds like hype. But the unit economics are not hypothetical. We operate across more than 30 markets. The pattern repeats. A city reaches density. Transport data compounds. Financial products emerge. The platform becomes infrastructure.

The mistake is looking at a map of countries and deciding where to deploy capital based on national population, national GDP, national Ease of Doing Business rankings. The signal is in the cities. It always has been. Abidjan told me that a decade ago. The numbers have been confirming it since.

Country GDP is a reporting abstraction. City GDP is where the work happens.