Modi's majority shrank and India's payment rails kept compounding. Those are two different kinds of power

Two curves moving apart

In June 2024 Narendra Modi was sworn in for a third term, but his party had lost its outright parliamentary majority and he formed a government at the head of the NDA coalition, dependent on allies including Chandrababu Naidu's TDP and Nitish Kumar's JD(U). Over broadly the same period, the Unified Payments Interface reached roughly twenty billion transactions a month worth about twenty-five lakh crore rupees. One measure of his power fell and the other kept compounding.

Coverage of any long-serving leader tends to collapse into one question about whether they are getting stronger or weaker, and the answer is usually taken from the last election. That framing misses the more useful distinction. Electoral power is the ability to pass things now and it is genuinely diminished by a coalition. Infrastructural power is what has already been built into the running of the country, and it does not consult the parliamentary arithmetic.

The numbers on each side

On the political side, the third term is a coalition government with the outright majority gone and allies holding real leverage over legislation. On the infrastructural side, UPI was running at roughly twenty billion transactions a month worth around twenty-five lakh crore rupees as of late 2025, is live in nine countries, has been adopted by the United Arab Emirates, and is being launched in Malaysia, with Modi urging BRICS members to take it up.

Why rails outlast mandates

Infrastructure, once laid, runs regardless of who holds the mandate; a majority does not.

The reason the two diverge is that they decay at different rates. A parliamentary majority is revalued every election and can vanish in a single cycle. A payment rail that a billion people and tens of millions of merchants use daily is not repealed by a new government, because the cost of removing it falls on the electorate rather than on the previous administration. Any successor inherits it, operates it, and takes credit for it. That asymmetry is why infrastructure is the more durable form of the two, and why it is a poor guide to a leader's current strength.

Separate the two kinds of power before judging either

Judge a government on what continues after it rather than on what it can currently pass.

Run the arithmetic and the scale becomes concrete. Twenty billion transactions a month is roughly 667 million a day, or about 7,700 every second, sustained. Twenty-five lakh crore rupees is 25 trillion rupees monthly, on the order of 300 trillion a year in flow. Those are utility-scale numbers, closer to a power grid than to a product, and utilities are not undone by elections. The coalition arithmetic, by contrast, is revalued every time an ally reconsiders.

What a single national rail concentrates

A single national payment rail concentrates operational and surveillance risk in one system.

Concentration is the cost of the achievement. A single rail carrying most of a country's retail payments is a single operational point of failure, and an outage that would be an inconvenience in a fragmented market is a national event in a consolidated one. It is also, by construction, a comprehensive record of economic behaviour, and the governance around who may query it matters more as the share of payments running through it approaches totality. Neither risk is hypothetical at this volume, and neither is addressed by the system working well.

Exporting a payment system

The export is the more interesting development. UPI operating in nine countries with the UAE adopting it and Malaysia launching turns a domestic system into an instrument of foreign policy. A country whose remittance and travel corridors run on your rails is integrated with you in a way no treaty produces, and the entity that sets the standard captures the tooling, the compliance layer and the talent pipeline around it. Encouraging BRICS members to adopt it is the same logic applied to a bloc.

The strongest argument against this

The reasonable objection is that this credits a government for something largely built by institutions and technologists rather than by a leader, and uses infrastructure to soften a genuine electoral verdict. Voters removed a majority; describing that as one of two power curves rather than as a defeat is the kind of framing that treats an election result as a technicality.

The objection is right that the electoral result was a real verdict and should not be softened, and the argument here does not soften it — it holds that the coalition constraint is severe and that legislation is genuinely harder now. What it adds is that the two measures answer different questions. Whether a leader can pass a bill next month and whether their term will still be shaping the country in twenty years are separate, and the second is not adjudicated by the first. Both readings are needed; neither substitutes for the other.

Seven thousand transactions a second is a utility, not a product

Dividing twenty billion monthly transactions down to the second gives roughly 7,700, sustained around the clock, every day. That is a throughput figure in the same class as a national grid or a telecoms backbone rather than a consumer application.

The distinction is not rhetorical. Products compete for users and can lose them; utilities are assumed, and their absence is what gets noticed. Once a payment system crosses into the second category, the relevant risk stops being competition and becomes reliability.

It also changes who the stakeholders are. At this volume the system's users include people who have never chosen it, because the merchant they buy from has, and that is the point at which participation stops being voluntary in any practical sense.

Exporting rails is cheaper than exporting influence

Nine countries live, the UAE adopted, Malaysia launching. Each of those is a corridor where remittances, travel spending and merchant settlement begin running on Indian-defined standards.

This is a familiar pattern in a new domain. Whoever sets a widely adopted technical standard captures the surrounding layer — the compliance tooling, the integration vendors, the trained engineers — without having to buy any of it. The standard does the work that money would otherwise have to do.

Compared with conventional instruments of influence, it is remarkably cheap. Aid must be renewed annually, treaties can be exited and military presence is expensive and resented. A payment rail becomes load-bearing quietly, and the cost of removing it rises every year it operates.

What the coalition actually constrains

It would be wrong to treat the coalition as a formality. Allies with pivotal seat counts have real leverage over contentious legislation, and the practical effect is that the most divisive items become expensive to attempt.

That is a genuine and material reduction in a specific capability: the ability to pass things quickly and unilaterally. Any analysis that skips past it in favour of the infrastructure story is doing the thing this piece is arguing against, in the opposite direction.

The accurate summary is narrow. Legislating is harder; operating what already exists is not. Those two facts sit side by side without resolving into a single verdict, which is uncomfortable and is usually why one of them gets dropped.

A caveat on the numbers

The transaction figures used here are the widely reported monthly totals as of late 2025, and volumes of this kind are revised and grow. They should be read as an order of magnitude rather than a current reading.

The per-second and per-day figures are derived by division from the monthly totals, not published as such, and they assume even distribution across the month, which real payment traffic never has. Peaks are considerably higher than the average implies.

Nothing here asserts a share of Indian retail payments, because credible estimates vary with definition. The argument rests on absolute scale and international adoption, both of which are reported directly.

What transfers outside government

The transferable question is the one most organisations never ask about their own work: what would still be running here if I left tomorrow, and what would stop.

Almost everything that depends on the founder's authority, relationships or continued attention belongs in the second category, however impressive it currently looks. Anything embedded in a process, a standard or a system that others now depend on belongs in the first.

The uncomfortable implication is that the most visible achievements are usually the least durable, because visibility tends to track personal involvement. The parts of the work that will outlast you are typically the ones nobody associates with you, and the trade between recognition and permanence is real.

How to tell durable power from electoral power

Ask which of a leader's achievements would survive their successor, and treat only those as durable.

List what a leader has actually changed and sort it into two columns: things that require continuing political authority, and things that now run on their own. Ask of each item in the second column what removing it would cost the public, because that cost is what protects it from a successor. Check whether the system has been exported or embedded elsewhere, since external adoption raises the removal cost further. Then look at what the concentration created, because durable infrastructure concentrates risk as reliably as it concentrates benefit.

The test is the first government that did not build it

Digital public infrastructure has never in India been operated by an administration that did not create it. The real measurement of durability will come whenever a different coalition inherits the system: whether it is maintained, renamed, restructured or quietly degraded. Until that has happened at least once, the durability argument rests on the logic of switching costs rather than on evidence.

Reassess on monthly transaction and value data rather than announcements, and treat the count of countries where the system is live as the more meaningful figure, since domestic growth eventually saturates while external adoption does not.

Editorial conclusion

A shrunken majority and a payment system at twenty billion transactions a month are both true, and reading either alone produces a wrong answer. The election imposed a real constraint on what can be passed. The rails impose a different kind of fact on what any future government inherits. The general lesson is that leaders are usually judged on the power that is easiest to observe, and that the power which lasts is almost always the part that has stopped being political and become plumbing.