Blitz Bureau
NEW DELHI: The headline numbers about UPI are about size. The number that explains it is about smallness — and it is the reason no other country has copied the model successfully.
In FY 2025-26 the Unified Payments Interface carried 24,161.69 crore transactions worth about ₹314 lakh crore, on the Finance Ministry’s record. This desk’s computation puts the daily average at 66.2 crore transactions and the average transaction at about ₹1,300. Some 703 banks were live on the network by March 2026, against 21 at launch in April 2016, and the International Monetary Fund has recognised UPI as the world’s largest real-time payment system by volume, at roughly 49 per cent of global real-time transactions.
Those are the size numbers, and they are impressive without being explanatory. The explanatory number sits further down the same record: of person-to-merchant payments, 86 per cent are below ₹500. Of person-to-person transfers, 59 per cent are below ₹500. The system that processes almost half the world’s instant payments is, in the overwhelming majority of its transactions, moving amounts that would previously have been settled with a hundred-rupee note and change.
The small-ticket system. Person-to-merchant payments account for 63 per cent of UPI’s volume; person-to-person transfers account for 71 per cent of its value. The two halves of the network are doing different jobs.
A payment system that is free at the point of use and profitable nowhere is a public good. The question a decade in is who pays to keep it running.
At a Glance
• FY 2025-26 volume: 24,161.69 crore transactions
• FY 2025-26 value: about ₹314 lakh crore
• Average transaction (this desk’s computation): about ₹1,300
• Daily average: 66.2 crore transactions
• Banks live: 703 as on March 2026, from 21 at launch
• Split: P2M is 63% of volume; P2P is 71% of value
• Small-ticket: 86% of P2M payments are under ₹500
• Global share: about 49% of world real-time payment volume
Read the volume and value splits together and the network resolves into two systems sharing one rail. Person-to-merchant traffic is 63 per cent of volume but a minority of value: it is the tea stall, the auto fare, the vegetable seller, the chemist. Person-to-person traffic is a minority of volume but 71 per cent of value: it is rent, school fees, the transfer to a parent, the settlement between traders. These have entirely different economics. High-volume, low-value merchant traffic imposes a real per-transaction cost on banks and payment providers and generates almost no float. High-value personal transfers cost the same to process and carry the same zero merchant discount rate.
This is the structural question that a decade of success has produced, and it is a question about durability rather than about failure. Digital public infrastructure of this kind is deliberately built as a utility — free at the point of use, interoperable by mandate, open to any licensed participant. That design is precisely why it reached 703 banks and eight foreign jurisdictions rather than fragmenting into walled gardens, and it is why the IMF ranks it first in the world. But a utility still has an operating cost, and at 66 crore transactions a day that cost is substantial and falls on the participants rather than the users.
The constructive path is already visible in outline and deserves to be completed deliberately rather than under pressure. Credit on UPI — overdraft lines, credit-card linkage, small-ticket lending against transaction history — monetises the rail without charging the small merchant, because it converts a payment record into a lending signal. Cross-border acceptance, live in eight jurisdictions and expanding, adds remittance and travel flows priced differently from domestic retail. And the 86-per-cent-under-₹500 figure is itself the strongest argument for keeping merchant payments free: a system that has brought the smallest traders in the country into a formal transaction record has produced a public asset whose value is not measured in fee income. The task of the next decade is to fund the rail from what it enables, not from what it moves.
