Around 2012, there was a serious push from the West to move the world off cash and onto electronic rails. The Better Than Cash Alliance was launched at the United Nations that September, with Visa among its many founders. It was also around this time that the world started seeing the immense merit in electronic payments. Four or five networks already carried almost everything that moved electronically, granting them the power to set the price of a payment and, if it came to it, to decide whether a payment happened at all.
That power did not stay hypothetical. Iranian banks were removed from SWIFT in 2012 and again in 2018; and in March 2022, Visa and Mastercard suspended their Russian operations within days. This is a crucial reminder that payments, like most infrastructure everyone is obliged to use, have always been instruments of political control. India’s position was that a country of its size could not sit on somebody else’s switch, which is why UPI was never meant to be a private protocol.
The three architectures of UPI
The thinking consolidated at the highest level early on. In 2009, the RBI’s Vision 2009-12 asked for vendor neutrality: an open, interoperable architecture, ensuring the country’s payment infrastructure would never depend on a single provider. UPI in 2016 was the natural extension of that instruction. It was India’s first serious attempt at letting private companies innovate on top of a public good, supported by three distinct architectures.
A decade of success has obscured how it was built: three architectures held it up. The policy architecture treated frictionless payment as a public good. A growing economy needs money to move at near-zero cost, and no private operator would price it there. The market architecture assumed that if payments were free and instant, companies would earn above the rail rather than on it. Credit, insurance, investments sold to customers acquired through it, etc. The technical architecture made both possible: it separated the custodian of money from the collector of consent. The bank held the account, the app took the authorisation, and let anyone pay from any app to any app. Unlike Account Aggregator, which is double-blind, UPI deliberately let the app see the transaction, leaving it a business built on data the rail generated free.
Shift from protocol to product
The business model worked better than anyone expected. FY26 closed at 24,162 crore transactions worth Rs 314 lakh crore across 554.9 million users. To put this in perspective, Mastercard switched about 175 billion transactions worldwide that year; UPI moved 242 billion inside one country.
However, the conversation around UPI pricing reveals that its foundational digital architectures are under strain. UPI has stopped being thought of as a protocol and started being run as a product. In FY25, the National Payments Corporation of India earned Rs 3,270 crore and kept Rs 1,552 crore as surplus. A true protocol would have reinvested that surplus back into the rail — into resilience against the inevitable outages a system of this size faces. Earning surplus directly from the rail proves it has changed from something the country runs into something someone sells.
Global adoption vs vendor specifications
What this commercialisation has cost us is easiest to see abroad. Twelve UPI linkages are live across 10 countries, but only three carry person-to-person payments; the rest exist so Indian tourists can pay foreign merchants. In FY25, this moved a mere `258 crore — a fraction of what UPI moves domestically in a day — and still settles through correspondent banks.
Compare this to MOSIP, the open-source identity platform built at IIIT Bangalore on the ideas behind Aadhaar. It is engaged with 31 countries and has 13 national roll-outs because countries simply download it and run it themselves. Or compare Pix, which Brazil open-sourced in 2022 and took into Argentina by March 2026. Governments adopt specifications, not foreign-controlled vendors. Had the policy architecture held, UPI would have travelled like MOSIP, as something other countries could own.
Flaws in the pricing debate
The same logic has come home. On August 6, Parliament amended the Section that made UPI free by law. The banks’ complaint is real: they carry the load on their core systems, and `2,196 crore in incentives does not cover it. Yet, this is not the full ledger. Districts where UPI grew fastest saw consumer durable lending compound ten times faster, and the transaction data making cash-flow lending possible sits on the banks’ own servers. Payment apps acquired hundreds of millions of customers for free — something no card network would have provided.
The belief that a cut of every transaction is the only way to fund UPI is flawed. Ad valorem pricing taxes the payment while ignoring where economic value actually accrues. Frictionless, programmable payments generate value well beyond any transfer fee. Distinct services — like IPO fund blocking, scheduled subscriptions, and margined futures — have measurable provision costs and can be priced without touching the transfer itself.
Preserving the public good
The efficient design prices the layer above and leaves the rail free. That the argument has devolved to a percentage of the transfer reveals a structural failure: the layer meant to earn was never built. A product reaches for the transfer fee because it is the only thing left to sell.
The question for the government is not what the rate should be, but what the rate is for. Europe drew this line in its Interchange Fee Regulation: whoever sets the standard cannot also price access to it. India must publish the UPI specification under an open licence with a working implementation, and hand it to a body that does not also run the switch. The only way past Brazil’s head start is to return to the design principles UPI was built with. We built the best argument in the world that a public good can carry a private industry on its back; it would be a strange thing to trade away.
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.