The National Payments Corporation of India announced last week that UPI processed 15.2 billion transactions in November 2025, worth about ₹23 trillion. It’s a staggering number—more than the entire GDP of most countries flowing through a single payment rail every month. And yet the institutions that make it possible are quietly bleeding money.

This is the part of the UPI success story nobody wants to talk about publicly. While NPCI celebrates growth metrics and the government points to India Stack as a global model, the banks and fintechs actually processing these transactions can’t figure out how to make the economics work.

The Problem With Free

UPI’s zero-merchant-discount-rate policy—meaning merchants pay nothing to accept UPI payments—was brilliant for adoption. When you remove friction from the merchant side, you remove the biggest barrier to digital payment acceptance. Small shops that would never pay 2% to accept cards happily display UPI QR codes because it costs them nothing.

But someone has to pay for the infrastructure. Payment systems don’t run on goodwill. There are servers to maintain, fraud detection systems to operate, customer support teams to staff, and regulations to comply with. In most countries, merchants fund these costs through interchange fees. In India, the government decided that would slow adoption, so they effectively banned charging merchants.

The result is that banks issuing UPI-enabled accounts bear all the costs with no corresponding revenue. They process billions of transactions and earn exactly nothing from them. Some banks tried to impose charges—remember when HDFC briefly floated the idea of fees for UPI transactions above a certain limit? The backlash was immediate and fierce. They backed down within 48 hours.

The Hidden Cross-Subsidy

Here’s how banks are surviving: they’re cross-subsidising UPI losses with profits from other products. Your home loan interest rate includes a premium to cover the fact that the bank loses money every time you pay your electricity bill via UPI. Your credit card annual fee funds the infrastructure that lets you split restaurant bills using PhonePe.

This works as long as banks have profitable product lines to subsidise the unprofitable ones. But competition is eroding margins everywhere. Mortgage rates are at historic lows. Credit card interchange is under regulatory pressure. Digital-only banks with lower cost structures are forcing traditional banks to cut fees.

At some point, the cross-subsidy model breaks. We’re not there yet, but the runway is getting shorter.

The Merchant Angle Everyone Misses

While merchants don’t pay explicit fees for UPI, they’re paying in other ways. Settlement times matter—cash is instant, UPI takes T+1 or longer depending on your bank. For a small business operating on thin margins, that working capital delay has a real cost.

More importantly, merchants can’t pass UPI costs to customers through pricing because UPI has become the default expectation. Twenty years ago, if you accepted credit cards, you could maintain slightly higher prices to cover the merchant discount rate and customers understood. Now, if you price in any UPI-related costs, you’re just seen as expensive compared to the shop next door.

This creates a race to the bottom where the payment system’s convenience for customers comes at the direct expense of merchant margins. It’s sustainable for large retailers with diversified revenue. For a kirana store doing ₹15,000 in daily UPI transactions, it’s another squeeze on already thin profits.

What Happens Next

The most likely outcome is gradual, quiet changes to UPI economics. We’re already seeing it—banks introducing minimum balance requirements that weren’t there before, reducing cashback on UPI transactions, pushing customers toward credit cards (where they earn interchange) instead of debit cards and bank accounts.

NPCI will probably introduce tiered pricing eventually. Not enough to kill adoption, but enough to stop the bleeding. Maybe merchants above a certain transaction volume pay a nominal fee. Maybe high-value transactions carry a small charge. These changes will be framed as “ensuring system sustainability” rather than admitting the current model doesn’t work.

The alternative is that a handful of large banks decide UPI is a strategic loss leader and everyone else becomes a dumb pipe—processing transactions for customers who maintain relationships with fintech apps rather than banks. That’s already happening with PhonePe and Google Pay dominating consumer-facing UPI volume while banks become invisible infrastructure.

The Global Implications

Other countries are watching UPI closely. Singapore, Thailand, and several African nations are building similar instant payment systems. They’re copying the technology but they’re not copying the economics, because they’ve seen how unsustainable zero-MDR becomes at scale.

Brazil’s PIX system, which processed 42 billion transactions in 2024, charges merchants for certain transaction types. It’s still far cheaper than cards, but it’s not free. That’s probably the smarter model—cheap enough to drive adoption, expensive enough to sustain operations.

India’s insistence on free-for-merchants was a political choice, not an economic one. It worked for adoption, but now we’re in the awkward phase where everyone knows the economics don’t work and nobody wants to be the first to say it publicly.

The Uncomfortable Math

Fifteen billion transactions per month is phenomenal adoption. It proves that Indians will embrace digital payments when the friction is low enough. But volume without viable economics just creates a bigger problem down the road.

Banks can’t run payment infrastructure as a permanent loss leader. At some point, either UPI introduces sustainable revenue models or the quality of service degrades as cost-cutting becomes the only lever banks have left.

We’re watching in real-time to see which way this goes.