Every power that mattered was built by a small circle of capable people and a state ambitious enough to use them. The difference now is that those people choose where to go.
Look behind any power that mattered and you find two things: a state with ambition, and a small number of people capable of carrying it out. Money, territory and armies followed from those two, rarely the other way round.
A nation's success is roughly proportional to the talent it can harness, multiplied by the ambition of the state that directs it.
What has changed is that talent now chooses. It crosses borders in weeks, not generations, and settles wherever it is trained, hired, trusted and used best. So the question for every state is no longer how much talent it has. It is how much it can keep, and what it does with it.
Sovereignty is possible only with assistance. A single wheel can never move.
Kautilya — Arthashastra, Book I (tr. R. Shamasastry, 1915)The states that rose fastest were the ones that collected minds. Persia's Sasanian kings took in the scholars Justinian expelled from Athens. The Abbasids paid translators at Baghdad's House of Wisdom to bring Greek and Sanskrit learning into Arabic. Chola temple colleges flourished in the same decades Chola fleets crossed the Bay of Bengal. Britain ran a subcontinent with barely a thousand Oxbridge-trained civil servants. The Birbal stories are mostly legend, but their moral holds: a king is only as wise as the cleverest person he will listen to.
The failure mode is a court that stops telling the truth. Saudi Arabia's The Line was sold as a 170-kilometre city for nine million people; by 2026 work had been halted with a sliver of it built. The problem was not simply a shortage of capital. It was whether capital was being directed toward plans that could survive contact with engineering, economics and execution, and my reading is that the court never forced that question. The same fund has since pivoted to ports and data centres, which is what a court that finally speaks up looks like.
Talent enters a country through the classroom or the job, and leaves when it stops trusting what comes next. Some states are trainers, like China and India. Some are attractors, like the UAE, where only about one resident in ten is a citizen. The United States has done both for a century, and immigrants founded more than half of its billion-dollar companies. Singapore's founder saw the logic early.
the more talent you have in a society, the better the society will grow
Lee Kuan Yew — Singapore's founding prime minister, 2011The fastest signal of that trust is where mobile money goes.
Britain, one of the richest countries on earth, is set to lose more millionaires than any nation on record after a decade of shifting tax rules. India's outflow, by contrast, has shrunk three years running. The pattern is not rich versus poor. It is predictable versus not.
No position is permanent. AI researchers moving to the United States, the greatest talent magnet in history, have fallen 89% since 2017, most of it in a single year. And for every foreign student enrolled in an Indian university, more than twenty Indians are studying abroad; the chief executives of Google, Microsoft, Adobe and IBM were all schooled in India and made their careers, and their citizenship, in America. A trainer that cannot retain is subsidising its rivals.
Eight of the world's ten best universities are American or British, and that is not an accident of history. When Sputnik flew in 1957, Washington answered within a year with the National Defense Education Act. It understood it was losing a talent race before it was losing a space race.
India is running the experiment in reverse. It has more ranked universities than ever, but none in the global top 100 for academic reputation, and education has been quietly losing its place in the budget.
The target India set itself in 1966 was 6% of GDP. It spends roughly half that, and its share of both Union and state budgets is lower than a decade ago. That is how a country ends up training brilliant undergraduates and exporting the graduate degree, the lab and the company that follows.
Countries do not dominate because they have great firms. They make ordinary firms great by giving them the world's talent, then defend the lead with state power.
America outspent China on AI roughly 23 to one, yet the gap between their best models is under 3%, because China now writes more of the world's AI research than anyone. Money buys compute. Talent density decides what it produces. Both governments know it: Trump flew to Beijing in May with Nvidia's Jensen Huang in tow, and by September the two states had agreed a standing AI dialogue. The talent sits in the companies. The strategy sits in the state.
The same logic runs through capital. Britain's empire was a chain of coaling stations and chokepoints; today China holds stakes in 129 overseas port projects, Dubai's DP World runs 82 terminals in over 40 countries, and India owns Haifa's port. A stake in a port buys leverage. It still takes people to run it.
Every ambitious government now runs the same playbook: nomad visas, tax holidays, land in a special zone. Malaysia did all of it for Balaji Srinivasan's Network School. Then, amid an election-season controversy in July 2026, local authorities revoked the operator's licence over its premises and signage, and the campus was ordered shut. Whatever the merits of the licensing case, the episode showed the clash at the heart of every talent strategy: incentives are written for founders, but enforcement answers to local politics. Kazakhstan signed the school the next day. Its offer will be tested the first time the campus becomes politically inconvenient.
| What the brochure says | What actually keeps people |
|---|---|
| Low or zero income tax | Rules that will not change retroactively |
| Fast-track visas | Visas that do not become political footballs |
| Subsidies and free zones | Officials who clear files rather than extract from them |
India shows the gap most sharply, because the money is already there. Incentive schemes and startup funds have multiplied. The ground a founder walks on has not moved as fast.
The system is such that it pays to be below the regulatory radar.
Gautam Chikermane — Observer Research Foundation, 2022India has begun closing that gap. PhonePe moved its domicile back in 2022, Groww followed in 2024, and Zepto later joined them, while companies including Razorpay and Meesho pursued similar moves. A 2024 reform created a fast-track route for inbound mergers, making future reverse flips substantially easier.
Capital follows friction removed, not money announced.
Removing friction is slow, unglamorous work. Its benefits often arrive beyond an electoral cycle, while its costs are immediate and concentrated. That gives every political system a version of the same problem: how to preserve attention for long-term state capacity when short-term demands are louder. Different systems solve that problem differently—and fail at it differently too. Britain has spent a decade protecting what its median voter already has; Indian states compete on pre-election handouts while education's share of their budgets falls. The Gulf's small, well-provided citizenries free their states to look outward, though that bargain has costs of its own, as the UAE learned under Iranian missiles this year. Ambition needs spare attention, and talent can tell which governments have it.
America built the greatest talent machine in history and is watching it slow. Britain makes world-class minds and loses the wealth they create. China trains at unmatched scale and still loses its rich. The UAE attracts brilliantly and is learning what openness costs. India trains superbly, has begun removing friction, and still exports too much of what it trains.
Talent no longer belongs to the country that produced it. It is on loan, renewed each year to whichever state it trusts most.
The next great power will be the country the world's best people choose to build in.