What the retail e-rupee does that UPI does not

The retail e-rupee (e₹-R) differs from a UPI payment in three ways. The money is a liability of the Reserve Bank of India (RBI), not of a commercial bank. A payment between two e₹ wallets settles at once without passing through either party’s bank account. And a sponsor such as a state government can attach conditions to it: where it can be spent, on what, and by when. The first two make little difference to someone buying groceries, and UPI is free for that payment. The RBI’s own case for the e₹ now rests on the third, and on cross-border payments.

The problem

The RBI launched the retail e₹ pilot on 1 December 2022 in four cities (pilot entry). By April 2026 it had 1 crore (10 million) users and about 15 crore transactions worth about ₹34,000 crore since launch, according to Deputy Governor T. Rabi Sankar (post-policy press conference, 8 April 2026).

The Unified Payments Interface (UPI), the instant payment system run by the National Payments Corporation of India (NPCI), carried 24,16,169 lakh transactions worth ₹314.2 lakh crore in 2025-26, about 86 per cent of all retail payment transactions in India (RBI Annual Report 2025-26, Table IX.1 and para IX.7). That is roughly 1,600 times as many transactions in one year as the e₹ handled in its first three years and four months. Governor Sanjay Malhotra put UPI at close to 490 million unique users and nearly 20 billion transactions a month in October 2025 (Global Fintech Fest keynote, 8 October 2025, para 5).

The e₹ in circulation fell over 2025-26, from ₹1,016.46 crore on 31 March 2025 to ₹771.66 crore on 31 March 2026 (RBI Annual Report 2025-26). In the same year the RBI recorded new features: users can send programmable money to each other, government agencies ran direct benefit transfer (DBT) pilots with programmable e₹, and offline payments are being tested (Wire). The question for the RBI, for the banks that distribute the e₹ and for the state governments now paying subsidies in it is what the e₹ adds in a country where UPI already covers instant, free retail payments.

Retail e-rupee in circulation at 31 March

Year-end value of e₹-R in circulation, ₹ crore.

Data

31 March	₹ crore
2023	5.7
2024	234.04
2025	1016.46
2026	771.66

Source: RBI annual reports, 2022-23 to 2025-26The Monsoon Ledger

How it works

Some terms first. A central bank digital currency (CBDC) is money issued by the central bank in digital form, a liability on its balance sheet like a banknote (RBI concept note, 7 October 2022, para 2.1). A bank deposit is money a commercial bank owes its customer. A wallet is the app account that holds e₹. Settlement is the transfer that discharges a payment, and finality is the point after which it cannot be reversed. A merchant discount rate (MDR) is the fee a merchant’s bank charges the merchant for accepting a digital payment.

Follow a customer paying a shop ₹200, first with UPI and then with e₹.

A UPI payment

The money before. The customer holds a savings account at a bank. The ₹200 is a deposit, a debt the bank owes the customer.
The payment. The customer scans the shop’s UPI QR code. The customer’s bank debits the account and the shop’s bank credits the shop’s account. At no point does the money stop being a commercial bank deposit: it moves from a claim on one bank to a claim on another.
Interbank settlement. The two banks now owe each other money, which they settle in central bank money on UPI’s settlement timelines rather than at the moment of payment (RBI e₹ FAQ, Q19).
The cost. The customer pays nothing. Under the framework announced on 15 September 2026, person-to-person UPI payments are free at any amount, and merchant payments up to ₹2,000 carry no MDR. A 0.4 per cent MDR, capped at ₹300, applies to merchant payments above ₹2,000, and banks are told to stop merchants passing it on to customers (Ministry of Finance, 15 September 2026). The shop in this example pays nothing either.

An e₹ payment

Loading the wallet. The customer opens an e₹ wallet with a bank or a non-bank in the pilot. Wallets are linked to the user’s savings account, so no separate know-your-customer (KYC) check is needed (RBI e₹ FAQ, Q11). The customer moves ₹200 from the bank account into the wallet, which works at any hour (Q9). The RBI creates e₹ and issues it to banks electronically, as it does with notes (Q6). The RBI’s concept note describes the balance sheet effect: the move shrinks the bank’s balance sheet, as when a customer withdraws cash at an ATM, and banks as a whole may end up holding fewer reserves at the RBI. The customer now holds an RBI liability instead of a bank one (concept note, para 7.5). The wallet shows the balance in note denominations, such as a ₹100 and two ₹50s (Q12).
The payment. The customer scans the shop’s CBDC QR code. The e₹ moves from the customer’s wallet to the shop’s wallet and settles at once, without passing through either party’s bank account (Q19). No bank owes another bank anything, so there is no interbank settlement step. If the customer has only a ₹500 token, the wallet pays ₹200 and keeps the change (Q13). There are no fees for using e₹ or e₹ wallets (Q15).
Paying at a UPI code instead. Most shops display a UPI QR code. An e₹ app can scan it, but the payment then settles on UPI’s timelines (Q19), so the instant, wallet-to-wallet settlement in step 2 applies only at CBDC QR codes.
Holding or redeeming. The shop can keep the e₹ or redeem it into its bank account at any hour (Q9). The e₹ earns no interest, because the RBI treats it as cash (Q17). Redemption reverses step 1: the bank’s deposits rise and the RBI’s e₹ liability falls.

A programmable e₹ payment

Programmability is where the e₹ and UPI differ most in practice. In Gujarat’s G-SAFAL scheme, the state pays livelihood assistance in programmable e₹ that beneficiaries can spend only on approved farm inputs within a geographic area (Governor’s keynote, 8 October 2025, para 21). In Gujarat, Puducherry and Chandigarh, public distribution system beneficiaries received food subsidies in programmable e₹, redeemable for eligible goods at fair price shops (RBI Annual Report 2025-26, para I.23).

The sponsor sets conditions. The government agency issues e₹ with conditions attached. The RBI lists expiry date, geographic location, merchant category code and the merchant’s UPI address as parameters (RBI e₹ FAQ, Q22).
The beneficiary receives it. Beneficiaries of a DBT scheme can sign up with a mobile number or Aadhaar authentication and do not need a savings account (Q11). The app shows which e₹ in the wallet is programmable and what conditions apply (Q24).
The beneficiary spends it. A payment to a merchant that meets the conditions settles. A payment that does not, such as one at an unapproved shop or after the expiry date, cannot be made with that e₹.

Why it is built this way

The RBI’s concept note of October 2022 sets out the design and the reasons for it. Five choices shape what the retail e₹ is today.

Banks distribute it. The RBI chose an indirect, two-tier model: it issues e₹, and banks and other providers onboard users, run wallets and carry out KYC and anti-money laundering checks (concept note, para 4.3.2). The note says the central bank has no comparative advantage in these customer-facing tasks, and that a direct model would disrupt the financial system and burden the RBI with onboarding.

It is a token, like a note. The note prefers a token-based retail CBDC because it is closer to cash: whoever holds the token is presumed to own it, and a token has a unique number that helps detect counterfeits (para 4.3.4). It also preferred fixed denominations so that users would trust the e₹ as they trust notes (para 4.4).

It pays no interest. An interest-bearing CBDC would draw money out of bank deposits and reduce banks’ ability to lend, the note argues. Cash pays no interest, so the e₹ does not either (para 4.3.3). The note also records the incentive problem this creates: banks “would restrain themselves from distributing CBDCs if they find it as a threat to their bank deposits” (para 4.3.3). Every rupee loaded into an e₹ wallet leaves the distributing bank’s deposits, as in step 1 above.

Privacy is managed, not full. The note proposes “anonymity for small value and traceable for high value”, similar to cash (para 4.3.5). The RBI has not published the anonymity threshold for the pilot. In practice, wallets are linked to KYC-verified savings accounts (RBI e₹ FAQ, Q11).

It should work offline and be programmable. The note says that only 825 million of India’s 1.4 billion people had internet access, so offline capability is needed for wide use, and it names double-spending as the risk to be managed with technical solutions and value limits (para 5.6). It describes programmability as tying money to an end use, giving agricultural credit spent only at input stores as an example (para 5.7). The RBI is testing offline payments over Near Field Communication (NFC) and SMS (RBI Annual Report 2025-26, para VI.54).

The concept note did not present the e₹ as a replacement for UPI. It describes India’s payment systems as available 24x7 with transaction costs “perhaps the lowest in the world” (para 3.2), says the e₹ is meant to complement existing payment systems rather than replace them (Summary), and says it should use existing infrastructure such as UPI so that merchants do not need a separate acceptance network (para 5.8). Its listed motivations start with the cost of cash: ₹4,984.80 crore spent on security printing in 2021-22 (para 3.3.1).

Four years later the RBI puts the case differently. In April 2026 Deputy Governor Rabi Sankar called programmability “the distinguishing feature of CBDC vis-a-vis other payment instruments”, and said a CBDC’s biggest advantage is in cross-border payments, which is why the RBI is in no hurry to launch it fully before other countries are ready (press conference, 8 April 2026). In August 2026 Deputy Governor Rohit Jain said the e₹ is used for real transactions and is called a pilot because the RBI is still testing technology and adoption (press conference, 5 August 2026).

In our markets

India is the only one of the three markets running a retail CBDC for the public. Nineteen banks and some non-banks offer e₹ wallets (RBI e₹ FAQ, Q21), and users can pay at both CBDC and UPI QR codes (Q2). The RBI has connected the two systems: the BHIM app, NPCI’s UPI app, can now find existing CBDC wallets (RBI Annual Report 2025-26, para VI.53). Gujarat, Puducherry, Chandigarh, Odisha and Andhra Pradesh have paid subsidies or benefits in e₹ (pilot entry). For 2026-27 the RBI plans a bilateral or multilateral cross-border CBDC pilot, and has held talks with the Monetary Authority of Singapore (MAS) and the Central Bank of the UAE (Wire).

Hong Kong tested a retail CBDC and stopped. The Hong Kong Monetary Authority (HKMA) ran two phases of e-HKD retail pilots covering programmability, offline payments and tokenised asset settlement. On 28 October 2025 it reported that the public perceived the e-HKD and tokenised bank deposits similarly, given their trust in Hong Kong’s banks, and moved its e-HKD work to wholesale payments (HKMA release; pilot entry).

Singapore decided against one. In March 2022 MAS, in a parliamentary reply by its deputy chairman Lawrence Wong, said the case for a retail CBDC in Singapore was not compelling, because financial inclusion is not a significant problem there and cheap, fast payments are already available through bank-based systems such as FAST and PayNow (MAS parliamentary reply, 2 March 2022). MAS has tested programmable payments with purpose-bound money under Project Orchid (MAS: Project Orchid).

The case against

Add the features to UPI and bank accounts instead. Each of the e₹’s distinguishing features has a version on existing rails.

Purpose-bound payments. e-RUPI, launched in August 2021 and developed by NPCI on its UPI platform (PIB, 31 July 2021), is a prepaid voucher tied to a person and a purpose, sent by SMS or QR code. The government described it as needing no bank account, card or payment app, and working on basic phones (PIB, 6 August 2021). It was designed for the same use as programmable e₹: making sure a benefit is spent on what it was given for.
Offline payments. The RBI’s framework for small offline payments allows up to ₹500 per payment and ₹2,000 in total on a card, wallet or mobile device, without an additional authentication factor. For UPI Lite the limits are ₹1,000 and ₹5,000 (RBI circular RBI/2021-22/146, 3 January 2022, updated 4 December 2024).
Settlement in central bank money. The concept note argues that e₹ payments are final and remove settlement risk between banks (concept note, para 3.3.3). The same note lists UPI among systems built for instant payment settlement (para 1.5). The risk the e₹ removes sits between banks, not with the payer or payee; no RBI document says this, and it is an inference from how the two flows are described.

On this view, the e₹ adds a second set of wallets, QR codes and apps for a benefit that users cannot see, while each wallet load takes deposits out of the banks asked to distribute it.

The adoption record. The fall in e₹ circulation in 2025-26 came in the same year the RBI added user-to-user programmability. One crore (10 million) e₹ users in April 2026 compares with close to 490 million UPI users in October 2025. Hong Kong and Singapore reached the conclusion this objection points to. The RBI’s answer is that the e₹ is still a pilot while it tests technology and adoption (press conference, 5 August 2026).

The answer depends on whom the e₹ is meant to draw. Judged by everyday payments, the objection holds. Judged by the two uses the RBI now names, it has not yet been tested.

The first is payments where the payer needs to control how the money is spent. A government paying a subsidy wants it used for farm inputs or rations, and programmable e₹ lets it set that condition on the money itself: the merchant category, the area, the expiry date. State and union territory governments, including Gujarat, Odisha and Andhra Pradesh, have used it, which is the use the RBI now leads with. For these payers the comparison is with a benefit transfer to a bank account, which carries no conditions, not with a UPI payment at a shop. The weakness is e-RUPI, which already offers purpose-bound vouchers on UPI. The RBI has not published how much of the e₹ in circulation, or how many transactions, come from programmable government payments, nor any measure of leakage in subsidy schemes before and after programmable e₹, and there are no comparable figures for e-RUPI. No published document shows that programmable e₹ delivers subsidies better than e-RUPI; this reading rests on states choosing it.

The second is cross-border payments, where the RBI places a CBDC’s biggest advantage. A domestic adoption figure cannot test that. The RBI’s 2026-27 cross-border pilot will, if retail wallets take part. If cross-border settlement runs between banks in wholesale CBDC instead, the retail e₹ loses this second reason and stands on programmable payments alone.

What to watch

2026-27: the RBI’s bilateral or multilateral cross-border CBDC pilot, with Singapore and the UAE named in its annual report, and whether retail wallets take part.
The RBI’s choice between NFC and SMS for offline e₹, and the value limits it sets.
Late May 2027: the RBI’s 2026-27 annual report, with e₹ in circulation at 31 March 2027, after a year of falling balances.
Whether the e₹ stays free for merchants now that UPI merchant payments above ₹2,000 carry a 0.4 per cent MDR.
An RBI decision to drop the pilot label, which Deputy Governor Rohit Jain called a matter of name only in August 2026.
Any published anonymity threshold for e₹, the “managed anonymity” the concept note proposed.