Much of the debate about artificial intelligence (AI) is about chatbots and automation. That matters for individual firms, but the bigger opportunity is national. AI is not primarily a technology story. It is an economic one.
In 1965, financier Michael Milken sketched a formula for prosperity that still holds: P = Ft × (HC + SC + RA).
Prosperity is Financial Technology multiplying Human Capital, Social Capital and Real Assets-the skills of our people, the trust in our institutions, and the farms, factories, power and digital infrastructure they depend on.
Kenya’s own history proves the point. In the 1960s and 70s, the cooperative movement was our first great financial multiplier-mobilising rural savings, financing inputs, turning subsistence farming into commercial enterprise.
The second arrived in 2007. M-Pesa cut the cost of moving money and pulled millions into the formal economy; today more than 38 million users transact over Sh40 trillion a year. The technology multiplied a cash economy Kenya already had.
The next multiplier will not simply be AI. It will be an AI-powered financial ecosystem. A multiplier only works on what it can reach-and AI reaches all three terms. It sharpens human capital, giving a farmer or a graduate sharper information and sounder decisions.
It strengthens social capital, replacing collateral and paperwork with verified records that make trust cheap and fraud costly. And it lifts the return on real assets-more harvest from the same land, more output from the same factory.
That is not hypothetical. Consider Apollo Agriculture, founded in Kenya in 2016. Using satellite imagery, soil data, machine learning and mobile-money histories, it underwrites smallholder farmers who have no collateral and no credit record-the customers banks will not touch. It has financed more than 350,000 farmers, who produce on average 2.6 times more than their neighbours.
Kenya is full of such success: M-Kopa lends against device-payment histories, banks and digital lenders against their own data. But each builds its private system alone-Apollo had to write its own credit models because none existed-and none of it is portable. A spotless M-Kopa record cannot follow you to a bank.
The engines are built; what is missing is the shared track they can all run on, so a farmer’s or a shopkeeper’s verified record is credit everywhere, not just inside one company’s app.
India built exactly this: identity, instant payments and consented data-sharing that let any lender assess a loan from verified records in minutes.
The same holds for the hardware beneath the algorithms. When Microsoft and G42 from UAE proposed a $1 billion AI data centre at Olkaria, the plan stalled: at full build-out it would have drawn close to a gigawatt-about a third of the national grid-and President Ruto warned Kenya could not switch off homes to power servers. But that constraint is the opportunity in disguise.
Kenya’s grid is already more than 90 percent renewable, and geothermal is its single largest source.
The Rift Valley holds an estimated 10 gigawatts of geothermal potential-five times today’s peak demand-and only a fraction is tapped. Cheap, clean, round-the-clock power is exactly what AI computing needs and few countries can offer. Build that steam, and Kenya will not merely host the AI economy-it can power it.
A fair question is what this means for jobs, especially the middle-level knowledge work-loan officers, assessors, analysts-that automation touches first. Some roles will change, and retraining is a real obligation we should fund, not gloss over. But Kenya’s problem has never been a surplus of such workers; it is that too few people are reached by them.
AI multiplies scarce expertise rather than replacing saturated labour: one skilled assessor can oversee a system serving thousands, extending credit and advice to the smallholder and the trader who had neither. In a mature economy AI substitutes for people.
In ours, it extends them.
The task now is to connect what exists into shared rails: interoperable public databases, secure consented data-sharing, a widened sandbox for AI-driven lending, and the clean energy to match.
Kenya is not starting from zero-a National AI Strategy, digital identity and digitised tax collection are already under way. But the obvious objection is trust: Kenyans have seen State data projects stall, and know how easily access becomes a toll booth.
That is the argument for building this the M-Pesa way-government setting open, consented standards, the market building and competing on top. A rail anyone accredited can join on published terms has no gate to guard and no middleman to pay.
Cooperatives did it for agriculture. M-Pesa did it for inclusion. AI can drive the next-not by replacing people, but by turning their daily work into credit a bank can price fairly and competitively, then release with speed.