Who Pays for Free Payments?

India’s UPI fee debate exposes the economics behind national digital payment infrastructure

“Free payments” is one of the most powerful ideas in financial technology.

For consumers, it removes hesitation. For merchants, it lowers the barrier to digital acceptance. For governments, it can support financial inclusion, reduce dependence on cash and accelerate formal economic activity.

But free describes the customer experience, not the infrastructure underneath it.

Every digital payment still needs to be authenticated, routed, processed, settled, monitored and reconciled. The system needs cybersecurity, fraud controls, dispute handling, regulatory oversight, software development and enough capacity to remain available during periods of extraordinary demand.

Someone has to pay for all of that.

India’s latest legislative development brings this question into focus.

On 4 August 2026, the government introduced the Taxation and Other Laws (Amendment) Bill in the Lok Sabha. Among other changes, the Bill would amend the Payment and Settlement Systems Act so that the central government can determine which electronic payment methods must remain protected from charges.

The Bill does not itself introduce a merchant discount rate for UPI. It does not specify a fee, transaction threshold or implementation date. UPI therefore remains free under the current framework.

However, the amendment creates the legal flexibility for a different pricing structure to be introduced later.

That makes the debate larger than one proposed law.

It raises a fundamental payment infrastructure question: how should a system be funded once it becomes too important to depend indefinitely on promotional economics and public subsidy?

Zero pricing helped build the network

The absence of merchant charges played an important role in UPI’s expansion.

Since January 2020, banks and payment system providers have generally been prohibited from charging payers or recipients for prescribed UPI and RuPay debit-card transactions. This gave merchants a strong economic reason to accept account-to-account payments, while consumers could use them without calculating an additional cost.

The result is one of the largest digital payment networks in the world.

UPI had 554.9 million onboarded users as of June 2026. It processed more than 241 billion transactions worth ₹314.23 lakh crore during the 2025–26 financial year. In July 2026 alone, the system handled a record 23.66 billion transactions worth ₹29.88 lakh crore.

This scale would have been much harder to achieve if every small merchant had faced a meaningful acceptance fee from the beginning.

Zero pricing helped create network effects. More consumers attracted more merchants, and broader merchant acceptance made the system more useful to consumers.

The policy worked.

The harder question is whether the same funding model remains appropriate after the network has reached national scale.

Free at the checkout does not mean free to operate

A UPI payment may appear almost costless to the customer.

Behind the interface, several institutions still perform work.

Banks maintain customer accounts and authentication systems. Payment applications acquire and support users. NPCI operates shared infrastructure. Acquirers connect merchants. Compliance teams monitor transactions. Fraud systems analyse behaviour. Customer-service teams investigate failures, scams and incorrect transfers.

The network also needs continuous investment in:

• processing capacity and availability;
• cybersecurity and application security;
• fraud detection and transaction monitoring;
• dispute and exception management;
• merchant onboarding and technical support;
• regulatory compliance;
• integration with banks and payment providers;
• product development and international connectivity.

India has already introduced enhanced security requirements, risk-based transaction limits and dedicated information-security frameworks for the UPI ecosystem. These controls improve resilience, but they also require technology, people and operational investment.

Transaction growth therefore increases both the value of the infrastructure and the resources needed to maintain it.

Volume alone does not fund the network when the price per transaction remains zero.

The current model relies on subsidy and cross-subsidy

When merchants do not pay a transaction fee, the underlying costs do not disappear.

They are redistributed.

The government currently supports low-value UPI payments to small merchants through an incentive scheme. The 2026–27 budget allocated ₹2,000 crore to the programme, following a reported payout of ₹2,196.21 crore in the previous financial year.

A parliamentary finance committee argued in March that the absence of MDR was making the UPI ecosystem financially unsustainable. It also reported that earlier incentives covered only a limited proportion of the payment industry’s estimated costs.

Other costs may be absorbed by banks, fintechs or payment providers because UPI helps them acquire customers, distribute other products or protect their position in the market.

That can work during rapid expansion.

It becomes more difficult when providers are expected to improve resilience, prevent increasingly sophisticated fraud and serve hundreds of millions of users without a direct transaction revenue model.

A payment rail should not depend permanently on participants treating infrastructure as a loss leader.

There are only a few realistic funding models

A large payment system can generally be financed through some combination of three approaches.

The first is public funding. The government subsidises the network because low-cost payments are treated as economic infrastructure, similar to other public digital services.

The second is participant funding. Banks, PSPs and applications absorb the cost because the payment system supports their broader commercial relationships.

The third is transaction pricing. Merchants or other participants pay a transparent fee connected to the cost and value of processing the payment.

None of these models is automatically correct.

Public funding can support inclusion, but it creates a recurring fiscal commitment. Participant funding can encourage broad adoption, but it may favour companies large enough to absorb losses. Transaction fees create a clearer economic model, but poorly designed fees can discourage small merchants and push activity back toward cash.

The real policy question is not whether payments should be free or paid.

It is how the cost should be distributed without weakening access, competition or infrastructure quality.

A tiered model is more credible than a universal fee

A universal MDR on every UPI transaction would risk undermining one of the system’s greatest achievements: easy and inexpensive digital acceptance for small merchants.

The stronger argument is for differentiation.

Person-to-person transfers and low-value payments to small merchants could remain free or publicly supported. Larger businesses, which receive substantial commercial value from digital acceptance, could contribute more directly to infrastructure costs. Higher-value transactions could also be treated differently from everyday micro-payments.

This would preserve the inclusion objective while recognising that a large online marketplace and a neighbourhood street vendor do not have the same economics.

The proposed legislation does not yet establish such a structure. Reports indicate that no fee, rate or timeline has been decided. Any tiered model remains a future policy possibility rather than a current rule.

The details would determine whether the outcome is balanced.

A badly calibrated fee could create avoidance behaviour or encourage merchants to favour cash. A modest and transparent charge applied to larger commercial activity could provide more reliable funding without materially changing consumer behaviour.

UPI should not be priced exactly like cards

The MDR debate also requires a careful comparison with card payments.

Card fees pay for more than transaction routing. Depending on the product, the card model may include credit, fraud liability, chargebacks, rewards, dispute rights and international acceptance. Issuers, acquirers, networks and processors each perform separate functions within the transaction.

UPI is an account-to-account system with a different risk and service model.

Its fee should not simply copy card economics.

A payment charge should reflect the services actually provided, the risk carried by each participant and the operational value delivered to the merchant.

This distinction becomes especially important as UPI adds recurring payments, credit-linked products, international connectivity and more complex merchant services. The underlying rail may remain efficient, while services built around it justify separate commercial pricing.

The cleanest model may therefore be low-cost common infrastructure combined with competitively priced services above the rail.

Sustainable economics support competition

Zero pricing can appear highly competitive because merchants pay less.

But if providers cannot recover their costs, the long-term effect may be the opposite.

Smaller fintechs may struggle to compete with banks or technology companies that can subsidise payments from other business lines. Product investment may slow. Customer support may weaken. Market concentration may increase because only the largest platforms can afford to operate at scale without direct payment revenue.

A modest revenue model can sometimes support more competition than a formally free market.

The important conditions are transparency and fair access.

Infrastructure fees should not be used to protect incumbents or create artificial barriers. Participants should understand what they are paying for, and pricing should remain proportionate to transaction value, merchant size and the services delivered.

The objective is not to make digital payments expensive.

It is to prevent “free” from becoming another word for underfunded.

What other markets can learn from India

Countries building instant-payment systems often focus first on adoption.

That is understandable. A payment rail has limited value without users, merchants and participating financial institutions.

But the funding model should be considered from the beginning.

A system can use temporary subsidies to accelerate adoption. It can protect low-value transactions and vulnerable users. It can establish public infrastructure while allowing private providers to compete through products and services.

What it should avoid is leaving the long-term economics undefined.

Once a rail becomes essential, changing the pricing model becomes politically difficult. Merchants and consumers begin treating the original subsidy as a permanent entitlement, while providers remain responsible for an expanding set of security and operational obligations.

India’s debate demonstrates the importance of separating two objectives:

Payments should remain broadly accessible. Payment infrastructure should remain economically sustainable.

Those objectives are compatible, but only when the funding structure is designed deliberately.

What businesses should consider

For merchants, a potential UPI fee is not simply another processing cost.

The relevant comparison is the complete payment outcome.

Businesses should consider acceptance cost, conversion, settlement speed, fraud exposure, refunds, customer preference, reconciliation quality and the operational effort required to support each method.

A payment method with a small fee may still be more valuable if it produces cleaner data, faster settlement or fewer failures.

Likewise, a nominally free payment can be expensive if it creates manual investigations, unclear references or difficult refund processes.

Payment strategy should therefore be based on total operating economics, not only the headline transaction price.

As national instant-payment systems expand, businesses will increasingly need the ability to compare different rails and route payments according to cost, geography, transaction type and customer preference.

MetaNord’s view

At MetaNord, we see India’s UPI debate as a sign of infrastructure maturity.

The system has already proven that low-cost, interoperable account-to-account payments can reach national scale. The question now is how to preserve that accessibility while funding the technology, controls and operating capacity required to keep the network reliable.

A strong payment system needs sustainable economics.

That does not mean every user should pay. It does not mean small merchants should lose access to affordable digital payments. It means the cost of operating the rail must be visible and allocated deliberately.

The customer experience can remain simple.

The infrastructure behind it cannot be treated as free.