UPI & Digital Public Infrastructure – UPI, India’s real-time payment network, built by the state and now exported abroad.
- 22.6bn UPI transactions (March 2026, single month)
- $3.3trn transacted in FY26 (across 241.6bn payments)
- ~49% of global real-time payments (processed in India)
- 24 countries signed on (India Stack / DPI MoUs)
In April 2016 a new payment system processed 373 transactions in its first month. In March 2026 it processed 22.64 billion of them in thirty-one days.
Across FY2025–26 UPI carried 241.6 billion transactions: about 660 million a day, across 703 banks, and roughly 85% of all digital payments in India. Many are tiny, because the network charges the merchant nothing to accept them.
The interesting part is the ownership. UPI is public infrastructure the state built, opened on equal terms, then declined to monetise. Twenty-four countries have signed on to the blueprint; it is live for payments in eight.
The signal to watch: Whether a merchant fee comes back, the point where public infrastructure has to start covering its own costs.
Where we would take your group
- A vegetable stall with a QR code. A printed code, a two-rupee sale, no merchant fee.
- A fintech building on the stack. Founders who treat payments as a public utility.
- A briefing on digital public infrastructure. Why the state owns the rails and not the apps.
Questions for the debrief
- UPI charges merchants nothing. That is why it won, and also why it earns nothing. Who should pay for critical public infrastructure once adoption is no longer the problem?
- Payments standardised easily; commerce (ONDC) has not. What properties make a market protocol-ready, and which of your students’ industries fail that test?
- A single state-linked identity and payment layer is powerful in both directions. Where would you draw the line on consent, data and exclusion?
