Digital Euro vs e-CNY vs Digital Rupee: CBDC Architecture Compared

Three central banks are building retail digital cash, and they have made three different bets on what that cash should be. The ECB is designing a privacy-first, offline-capable public utility that banks distribute but do not own. The People’s Bank of China has just turned the e-CNY into interest-bearing bank deposit money. The Reserve Bank of India runs a token-style wallet that rides on top of UPI and is now being tested as a programmable subsidy rail. Any serious CBDC architecture analysis has to start from that divergence, because the design choices that look like technical details, such as who holds the liability, whether balances earn interest, and where the offline value lives, decide who the system serves.

This matters now because the policy clock moved in 2026: the European Parliament adopted its position in July, China’s new framework took effect on 1 January, and India’s pilots shifted from novelty payments to welfare delivery. This post compares the three programmes layer by layer, using only sourced facts, and flags what is still undecided.

What this covers: the two-tier distribution model, ledger and instrument design, offline payments, privacy tiers, programmability, and a decision matrix with failure modes. This is systems analysis, not financial or legal advice.

Context and Background

A central bank digital currency (CBDC) is a digital liability of the central bank denominated in the national unit of account. Retail CBDCs are meant for households and merchants, while wholesale CBDCs are meant for settlement between financial institutions. The three programmes compared here are retail systems, although India also runs a wholesale pilot and China connects its currency to cross-border experiments. Almost every serious design since 2020 adopted the same macro-shape, a two-tier model, in which the central bank issues and settles while licensed intermediaries onboard customers and run the wallets. The reason is practical: central banks do not want to operate customer support, anti-money-laundering (AML) onboarding, or fraud desks for millions of users.

The differences emerge below that shared shape. Each programme answers four questions differently: is the instrument a claim on the central bank or on a commercial bank; is it account-based, where identity gates every payment, or token-like, where possession carries value; can it work with no network; and how much of the transaction record does the issuer see. Those answers reflect each economy’s incumbent payment stack. Europe has card schemes dominated by non-European networks; Visa and Mastercard account for 61 percent of eurozone card payments according to reporting on the Parliament vote. China has two private super-apps, WeChat Pay and Alipay, which dominate mobile payments. India has UPI, a public-good instant payment rail that already processed 22.64 billion transactions in March 2026 according to one explainer citing NPCI data.

That last contrast is the key to understanding why the programmes diverge. A CBDC enters a market that already has a payment habit, and it has to offer something the incumbent does not. The digital euro offers European sovereignty and a legal-tender status across the euro area. The e-CNY offers a state-backed alternative to two private duopolists, which is why China keeps adjusting incentives. The digital rupee competes with a rail that is nearly free and extremely convenient, so its differentiator is programmability and offline use rather than raw payment speed.

For readers building payment infrastructure, there is a second connection worth noting. The messaging layer under all three is moving to ISO 20022 data models for wholesale and bank-facing flows, which we cover in our ISO 20022 migration payments architecture guide. The CBDC retail layer and the bank-facing layer have to reconcile, and richer structured data is what makes that reconciliation feasible. The ECB itself describes its project through its official digital euro pages, which are the primary source for the preparation timeline.

A caution on evidence quality. The ECB, PBoC, and RBI publish policy statements, but few publish detailed protocol specifications. Where this post gives technical detail, such as wallet tier limits or offline modes, it names the date and source of the detail, because several widely repeated figures describe the 2021 pilot designs and not the current rules.

Reference Architecture: The Tiered Model and Three Variations

All three programmes use a two-tier architecture in which the central bank settles and licensed intermediaries serve customers, but they differ in who owns the customer relationship, the legal nature of the balance, and the interoperability layer on top.

Figure 1: Shared two-tier skeleton with three distinct distribution choices for the digital euro, e-CNY and digital rupee.

Figure 1 shows the shared skeleton and the three branches. The long-description version: the central bank sits at the top as issuer and settlement authority; licensed intermediaries sit in the middle; users and merchants sit at the bottom; and each programme attaches its own distribution choice to the middle tier.

Digital euro: infrastructure provided by the ECB, relationships held by supervised intermediaries

In the digital euro design, the ECB provides the underlying infrastructure and commercial banks and payment service providers (PSPs) deliver the services that customers see. This follows reporting on the Parliament position and the stated ECB approach. Intermediaries are compensated, and the structure of that compensation remains, in the words of one report, among the most contentious issues. The Freshfields summary of the Council and Parliament positions describes a transitional model with euro-area uniform fee caps based on comparable payment means, followed by a later shift to a cost-plus-profit system, with disagreement on how long the transition may last: the Council wants a ten-year maximum while Parliament would allow it to continue if efficiency is shown.

The architectural consequence is that the digital euro is designed as a scheme, much like a card scheme or SEPA Instant, with a rulebook, certified PSPs, and a settlement layer. The ECB’s role is closer to a scheme operator with a monetary-liability twist. Pilot participation reflects this: reporting says 36 banks and PSPs were selected for a pilot starting in the second half of 2027.

e-CNY: operator banks and, since 2026, deposit money

The e-CNY has used a two-tier operator model since its pilots began: the PBoC issues, and designated operator institutions exchange, distribute, and manage wallets. What changed on 1 January 2026 is the legal nature of the balance. According to Xinhua’s report on the PBoC action plan, digital yuan held in commercial bank wallets is now treated as bank deposit liability, bearing interest in line with prevailing deposit rate regulations, with deposit insurance equivalent to ordinary deposits. Non-bank payment institutions must hold 100 percent reserves against the digital yuan they manage.

This is a major conceptual shift. The e-CNY moves from being a cash-like instrument toward what Xinhua calls a form of digital deposit money. Trivium China characterised it as a reclassification from a digital cash substitute to an interest-bearing deposit, and noted the aim of giving financial firms a reason to promote it against WeChat Pay and Alipay.

Digital rupee: a token wallet over bank apps and UPI

The RBI’s retail pilot launched on 1 December 2022 and its wholesale pilot on 1 November 2022. Retail wallets are provided by participating banks; one 2026 explainer counts 19 banks, while Wikipedia, citing earlier data, cites 17 banks as of March 2025. The e-rupee wallet sits next to, not inside, a bank account: the user downloads a bank app, registers, sets a PIN, and loads funds from a linked account. Several banks integrated UPI compatibility so that existing UPI virtual addresses and QR codes can be used to pay with the digital rupee without separate merchant onboarding.

That interoperability decision is the cleverest part of the Indian design. Instead of asking merchants to install a new acceptance method, the RBI let the digital rupee piggyback on the acceptance footprint UPI already created. It also means the customer experience at the point of sale is nearly indistinguishable from UPI, which shapes how adoption should be measured.

Why the distribution choice is the real architecture

Ledger technology gets most of the attention in CBDC discussions, but the distribution choice constrains more. Whoever owns the customer relationship owns the onboarding, the fraud liability, the data, and the margin. Europe chose supervised intermediaries under a regulated fee cap. China chose to move the balance onto bank balance sheets. India chose to hide the CBDC behind an existing front end. Each of those is a political-economy decision implemented as a system boundary.

Deeper Analysis: Ledger Design, Offline Value, Privacy and Programmability

The most useful way to compare the three systems is to separate four design layers: the legal instrument, the settlement and ledger model, the offline mechanism, and the privacy model. Each layer has trade-offs that the others cannot fully repair.

The legal instrument: central bank liability, deposit, or token

Figure 2 places each instrument by the balance sheet that carries it and shows that all three flows end at settlement at the central bank.

Figure 2: Instrument type by programme. The e-CNY bank wallet is now a deposit liability, while the digital euro and digital rupee remain central bank liabilities.

The digital euro is intended as a direct claim on the Eurosystem, legal tender alongside cash. Both the Council and Parliament positions, as summarised by Freshfields, recognise it as legal tender with exceptions based on entity type and size, and Parliament specifically exempts small and micro-enterprises from mandatory acceptance. Because it is a central bank liability, the euro area wants to prevent it from draining bank deposits in a crisis. That is the reason holding limits are a core design parameter. Both institutions agree caps are needed to prevent use as a store of value; they disagree on who sets them.

According to the PPC Land report on the ECON vote, individual holdings would be subject to EU-wide ceilings determined by the European Commission based on ECB recommendations, with a mandatory review every two years. Businesses could not hold digital euros except to accumulate incoming payments for up to 24 hours before transferring funds. The Council, by contrast, prefers a Council implementing decision with reinforced qualified majority. The actual ceiling has not been fixed in law. Secondary reporting has cited an ECB figure of 3,000 euros as an analytical reference for the holding limit; I could not verify that figure against an ECB primary document in this run, so treat it as reported and illustrative, not as a decision.

The e-CNY now sits at the opposite end. A bank wallet balance is a liability of the bank, backed by the deposit insurance regime and counted in the reserve requirement base, and it pays interest. This eliminates the disintermediation worry that drives European holding limits, because the money never leaves the banking system’s balance sheet. The price is that the e-CNY in bank wallets is no longer meaningfully “digital cash.” Non-bank payment institutions are held to a 100 percent reserve requirement, which makes a wallet provider a custodian of central bank reserves and not a credit creator.

The digital rupee is a token-style central bank liability: it is issued by the RBI, held in wallets provided by banks, and earns no interest, though it can be converted to bank deposits. This is by design. The RBI’s public positioning stresses substitution for cash and not competition with deposits. The absence of interest is itself a holding-limit substitute, since there is no incentive to park large sums.

Ledger and settlement: what is public and what is not

A frequent misconception is that a CBDC means a blockchain. None of the three retail programmes publicly commits to a permissionless ledger. Public documents describe two-tier distribution and settlement at the central bank, and details of the internal ledger technology vary in what has been disclosed. I could not verify a published ledger specification for any of the three in this run, so this post does not characterise their internal data stores. The RBI’s early concept note, summarised on Wikipedia, discussed token-based and account-based approaches and single-tier versus two-tier distribution, and the pilot adopted the two-tier model with a token-style retail wallet and an account-style wholesale design.

What can be said precisely is the shape of the settlement problem. Whatever the data store, a retail CBDC must provide finality for each payment, a way to reconcile intermediary positions with the central bank, a defence against double spending, and an audit trail for supervisors. Our work on feature store architecture and online-offline parity is unrelated in domain but shares one structural lesson: when two stores, one fast and local and one authoritative and central, must agree, the interesting engineering is in the reconciliation protocol and not in either store. CBDC offline value has the same shape, which is why the offline mechanism deserves separate treatment.

Offline payments: value held on a device

Offline payment is the CBDC feature with no equivalent in account-based fast payment systems. Figure 3 shows the generic protocol: a payer device with a secure element debits a local balance and signs a transfer; the merchant device verifies the signature and an anti-replay check; both sides sync later when a network is available.

Figure 3: Generic offline CBDC payment sequence. Exact protocols for each programme are not fully public; this is an illustrative pattern.

The digital euro takes the strictest version of this model. In the ECON position described by PPC Land, offline payments work through local storage devices without network connection, directly equivalent to carrying physical cash, and losing the device means losing the funds with no refund. Offline payments are free of transaction fees. Both Council and Parliament mandate online and offline versions, according to Freshfields, and the Council text details offline payment verification rules. The ECB also wants offline use to be the most private mode: both sides agree the digital euro should be highly privacy-preserving, especially offline.

India enabled offline transactions using near-field communication (NFC) by late 2024 per Wikipedia’s summary of RBI announcements, and a 2026 explainer describes NFC payments in low-internet areas. One news report on pilots in rural Odisha and Bihar mentions Bluetooth wallets with a small value limit; I did not verify that report’s specifics against an RBI source, so I do not repeat its limits here.

China’s 2021 e-CNY wallet specification described both soft wallets (mobile apps) and hard wallets (IC cards, wearables), which is the hardware basis for card-style offline use. I cannot confirm the current offline limits in this run.

The central engineering problem is double spending without a network. A secure element can enforce a counter and a balance locally, but if the secure element is compromised the attacker can spend the same value twice before any sync occurs. Every offline design therefore combines three mitigations: a hard cap on offline balances, tamper-resistant hardware or attested software, and a sync-time reconciliation that detects and penalises duplicates. The cap bounds the loss; the hardware raises the cost of attack; the reconciliation turns an attack into a recoverable event for the system, if not for the merchant who accepted the duplicate.

A short worked example shows why caps matter. Suppose, for illustration only, that an offline cap is 500 units and an attacker can clone a device and spend its balance at twenty merchants before reconciliation. The loss is 500 times 20, so 10,000 units, divided across twenty merchants who each took a 500-unit payment. If the cap is 50 units the same attack costs 1,000 units. The figures are invented to show scaling, not any programme’s real parameters, but the linear relationship between cap and exposure is why every programme limits offline holdings.

Privacy: three models, three different trust boundaries

Privacy is the layer where the programmes diverge most sharply and where the vocabulary is slipperiest. Three different things are called privacy: confidentiality from merchants, confidentiality from intermediaries, and confidentiality from the state.

The digital euro’s stated goal is that the ECB cannot directly identify users from payment data. The ECON position, as reported, incorporates privacy-by-design and privacy-by-default and mentions cryptographic techniques such as zero-knowledge proofs, with the ECB not accessing personal identification data. The Council and Parliament still disagree on which categories of personal data PSPs and central banks may process, with Parliament emphasising alignment with the General Data Protection Regulation (GDPR). In practice, online digital euro payments remain visible to the intermediary, which must run AML checks; the strongest privacy applies to offline payments, which behave like cash.

The e-CNY uses what the PBoC calls “controllable anonymity.” The 2021 wallet specification, reported by Forkast, defined a lowest tier activated with only a mobile phone number, with a single-transaction limit of 2,000 yuan, a daily limit of 5,000 yuan, and a balance limit of 10,000 yuan, and a fully verified tier linked to an ID card and bank account with a 50,000 yuan transaction limit, 100,000 yuan daily limit, and 500,000 yuan balance limit. These are 2021 figures; current limits may differ and I could not verify them in this run. The architectural point is the tiering principle: low-value anonymity for the user, with the central bank and operator able to see more as identity assurance and limits increase.

India’s model is similar in principle. Smaller sums remain anonymous while larger sums require self-disclosure for AML compliance, per Wikipedia’s summary. Tiering by value is the common answer to the tension between cash-like privacy and financial-crime controls, and all three programmes use some version of it. What differs is which actor, the central bank, the intermediary, or the state, can unmask a transaction, and under what legal process.

Programmability: purpose-bound money in India

Programmability means attaching conditions to money itself, and India has pushed this furthest. As of 30 August 2024, programmable features expanded to fuel, groceries, education, dining, healthcare, and travel, per Wikipedia’s summary. In 2026, a welfare pilot launched on 26 February in Puducherry for a food subsidy, and about ten welfare programmes were reported as active pilots, with a Pradhan Mantri Garib Kalyan Anna Yojana pilot using programmable tokens restricted to authorised foodgrain redemption, according to one explainer. That explainer also cites a welfare budget of roughly 80 billion US dollars threaded through the e-rupee, a figure I could not verify against a government source and which should be read as an aspiration or an estimate rather than current flow.

The ECB, by contrast, has explicitly framed the digital euro as general-purpose money and is cautious about programmable money in its design, though conditional payment is supported at the intermediary layer. I did not verify the current wording in this run, so treat that as a positioning summary and not a quote. The architectural lesson from India is that purpose-bound tokens move enforcement from the merchant’s point of sale to the issuance and redemption logic, which reduces leakage for subsidies but creates a new governance question: who decides which merchants and goods are eligible, and how is that list audited.

Programme Status and Decision Matrix as of October 2026

Figure 4 summarises the European timeline, which is the one still in flux. The e-CNY and digital rupee are already operating, so their question is scale and design evolution, while the digital euro’s question is whether and in what form legislation passes.

Figure 4: Digital euro path from preparation phase to earliest issuance. Trilogue outcomes are unknown as of this writing.

Digital euro: legislation first, infrastructure second

The ECB announced on 30 October 2025 that it would move to the next phase, with pilot exercises and initial transactions around mid-2027 and a technical readiness target for possible first issuance in 2029. The ECB stressed that the final decision on whether to issue a digital euro, and when, will only be made once the relevant legislation has been adopted. The Council adopted its mandate on 19 December 2025. The ECON committee voted 43 to 14 with one abstention on 23 June 2026, and the plenary adopted the Parliament position on 9 July 2026, reported as 416 in favour, 169 against, and 22 abstentions, approving trilogue negotiations with the Council under the Irish Presidency.

As of this run I did not find a report of a concluded trilogue deal, so the legal text, including the holding limit mechanism, the compensation model, and the treatment of legal persons, is not final. Parliament proposes a near-total prohibition on holdings by legal persons, while the Council suggests ECB discretion guided by a recital. The ECON text also envisages a rollout period of at least 24 months after authorisation. If those provisions survive, a 2029 first issuance depends on a quick agreement, because the rollout window consumes time after the law passes. That is my reading of the timeline arithmetic, not an ECB statement.

e-CNY: scale and the January 2026 reset

Xinhua reported 3.48 billion cumulative transactions worth 16.7 trillion yuan as of November 2025, which is around 2.37 trillion US dollars at the exchange rate used in that report. That figure is cumulative since pilots began and should not be read as a run rate. Since 1 January 2026 the framework treats bank-wallet balances as deposits, so any comparison of “e-CNY in circulation” before and after that date mixes two different definitions. The earlier statistics counted e-CNY as cash in circulation (M0); the new treatment moves bank-wallet balances into deposit money. I did not retrieve PBoC data post-reclassification in this run.

Digital rupee: pilots, scale, and the UPI shadow

One 2026 explainer reports 10 million users by April 2026 and about 3.6 billion US dollars of cumulative transaction volume since December 2022. Wikipedia reports 1,016 crore rupees in circulation, around 120 million US dollars, across 17 banks and 6 million users in March 2025, and notes that daily transactions peaked above 1 million in December 2023, then fell to around 100,000 by June 2024 following system stress tests. These sources differ in date and provenance, and I could not confirm any against an RBI release in this run, so the numbers should be treated as indicative. Even so, the ratio is informative: the same explainer cites 22.64 billion UPI transactions in a single month, March 2026. The digital rupee’s cumulative volume since 2022 is small next to monthly UPI flows, and that is the central adoption challenge.

Decision matrix

Dimension	Digital euro	e-CNY	Digital rupee
Status (Oct 2026)	Legislation in trilogue; ECB pilot planned second half 2027	Live; deposit-based framework since 1 Jan 2026	Live pilots; welfare programmable pilots
Liability	Eurosystem (central bank)	Bank deposit in bank wallets; non-bank wallets fully reserved	RBI (central bank)
Interest	Not stated in sources reviewed; holding limits used to contain demand	Yes, per deposit rate rules in bank wallets	No
Distribution	Supervised banks and PSPs, ECB infrastructure	Operator banks and institutions	Participating banks, UPI-compatible QR
Offline	Mandated in both positions; cash-like, no refund on loss	Hardware and soft wallet heritage; current limits unverified	NFC offline by late 2024 per secondary source
Privacy	High, ECB cannot identify users; strongest offline	Tiered, controllable anonymity	Small sums anonymous, larger need disclosure
Programmability	Not a headline feature	Not verified in this run	Purpose-bound tokens in welfare pilots
Cost to merchants	Fee caps proposed; offline free	Not verified in this run	Not verified in this run

Which design fits which problem

If the goal is sovereignty over a regional payment stack with strong privacy guarantees, the digital euro design is the most coherent, but it is also the slowest because it must pass through a legislative process involving 27 member states. If the goal is to push a state-backed instrument through banks that can compete with private super-apps, the e-CNY approach of paying interest and making wallets bank deposits is the most aggressive tool. If the goal is purpose-bound public spending on top of an existing instant payment rail, the Indian combination of a token wallet, UPI-compatible acceptance, and programmable redemption is the most practical.

A fourth consideration is agentic commerce. Software agents that pay on behalf of users need programmable, low-friction settlement with clear authorisation boundaries. Our analysis of agentic payments architecture for AI commerce argues that mandate and delegation models matter more than the rail itself. A CBDC with programmable redemption, as in India, maps cleanly onto mandate constraints, whereas a privacy-first offline-capable instrument, as in the euro design, fits human-present payments better. That is an architectural inference, not something any central bank has announced.

Trade-offs, Gotchas, and What Goes Wrong

Every CBDC design makes trade-offs that show up as failure modes in production. The first is disintermediation risk. A central bank liability that is as convenient as a bank deposit can pull funds out of banks in stress, accelerating a run. Holding limits, zero interest, or conversion to deposit status are the three available levers, and each programme pulled a different one: the euro uses caps, India uses zero interest, and China made the balance a deposit.

The second is the adoption trap. Merchants accept what customers carry, and customers carry what merchants accept. The Indian numbers illustrate it: a dedicated wallet with UPI-compatible acceptance still shows cumulative volume that is tiny against monthly UPI flows. Mandatory acceptance, which the European legislation proposes with exemptions for small enterprises, is a legal fix for a market problem, and it raises compliance cost for merchants.

The third is offline fraud. As the worked example above shows, exposure scales with the offline cap and the number of offline merchants an attacker can reach before reconciliation. The strictest design, no refund on loss as in the ECON text, also raises consumer-protection questions, because a lost phone becomes a lost wallet with no recourse.

The fourth is privacy theatre. Claims of anonymity tend to hold at the edge, such as offline small payments, and weaken at the intermediary, where AML obligations apply. A reader should always ask who can link an identity to a payment, and under what legal process, before accepting a privacy claim.

The fifth is definitional drift in the data. China’s January 2026 reclassification and India’s differing counts of participating banks show how statistics change meaning. Any dashboard that tracks “CBDC in circulation” needs a metadata column for the definition in force on each date, or time series will silently break. The same discipline applies to pipelines that ingest from central bank APIs and data feeds, where schema versioning is a first-class requirement.

The sixth is political fragmentation. If each region ships a different CBDC with different privacy, limit, and interest rules, cross-border interoperability becomes an exercise in mapping three legal instruments. India is reported to be planning bilateral and multilateral cross-border CBDC pilots for 2026 to 2027, including an MOU with the Monetary Authority of Singapore, per one explainer; I did not verify this against an RBI statement.

Finally, there is the quiet failure of governance by pilot. A pilot’s parameters, such as limits and eligible merchants, are policy choices that tend to become permanent. Designers should write down which parameters are experimental and which are commitments.

Practical Recommendations

For payment architects and fintech engineers, the useful takeaway is to design for CBDC as a set of pluggable instrument types and not as a single new rail. Treat the instrument type, whether central bank liability, deposit, or token, as a first-class attribute on every balance record, because compliance, reserve, and interest treatment follow from it.

For teams preparing to integrate the digital euro, wait for the trilogue outcome before hardcoding limits, since the holding ceiling, the compensation model, and the treatment of legal persons are all unsettled. Build abstraction around limits and fees so that a delegated act or implementing decision changes configuration, not code. For teams working with e-CNY, handle the post-January 2026 split between bank-wallet deposit balances and non-bank wallet balances as separate reserve and risk categories. For teams in India, treat UPI-compatible QR acceptance as the default integration path and programmability as an opt-in feature for specific use cases.

A short checklist:

• Record instrument type, legal liability, and interest treatment for every CBDC balance.
• Externalise holding limits, fee caps, and offline caps as versioned configuration.
• Model offline exposure as cap times reachable merchants before reconciliation.
• Separate identity assurance tiers from payment limits in your data model.
• Version every statistic with the definition in force on that date.
• Re-verify programme status against primary sources before each quarterly planning cycle.

Non-advisory note: this article is systems analysis and does not constitute financial or legal advice.

Frequently Asked Questions

What is the difference between the digital euro, e-CNY and digital rupee?

All three are retail central bank digital currencies built on a two-tier model, but they differ in legal nature and status. The digital euro is a planned central bank liability still moving through EU legislation. The e-CNY is live and, since 1 January 2026, is treated as interest-bearing deposit money in bank wallets. The digital rupee is a live pilot token-style wallet that earns no interest and is increasingly tested for programmable subsidies.

Will the digital euro replace cash?

No, according to the legislative positions summarised in the sources reviewed. Both Council and Parliament treat the digital euro as legal tender alongside cash, and the Parliament bill includes provisions to preserve cash as a payment instrument. The ECB also frames it as a complement. Mandatory acceptance applies with exceptions, and Parliament exempts small and micro-enterprises. The final legal text depends on the ongoing trilogue negotiations, so details could still change.

Does the e-CNY pay interest?

Since 1 January 2026, yes for balances in commercial bank wallets. Xinhua reports that banks must pay interest on digital yuan wallet balances in accordance with prevailing deposit rate regulations, and that these balances carry deposit insurance equivalent to ordinary deposits. Non-bank payment institutions must hold 100 percent reserves against the digital yuan they manage. I could not verify the specific rates paid, which follow the deposit rate rules rather than a separate e-CNY rate.

Can the digital euro or digital rupee work offline?

Yes in both designs. The European Parliament and Council both mandate online and offline versions of the digital euro, and the Parliament position describes offline payment as equivalent to carrying physical cash, with no refund if the device is lost. The digital rupee supports NFC-based offline payments according to secondary reports from late 2024 onward. Exact offline limits and protocols are not fully public, so verify them against central bank documentation.

Is a CBDC the same as a stablecoin or cryptocurrency?

No. A CBDC is a direct or indirect liability of the central bank in the national currency, with legal tender status in the digital euro case and with the central bank controlling issuance and rules. Stablecoins are issued by private entities and backed by reserves, and cryptocurrencies like Bitcoin have no issuer. A CBDC does not require a public blockchain, and none of the three programmes discussed here publicly commits to one for retail.

When will the digital euro launch?

The ECB has said it could be ready for first issuance as early as 2029, with pilot exercises beginning around the second half of 2027, but only after legislation is adopted. The Parliament adopted its position on 9 July 2026 and trilogue negotiations with the Council are the next step. The ECON text envisages a rollout period of at least 24 months after authorisation, so the timeline is tight and depends on the pace of agreement.

Further Reading

• Agentic payments architecture for AI commerce: how mandates and delegated authority shape payments that software agents initiate.
• ISO 20022 migration and payments architecture: the structured messaging layer that CBDC settlement flows reconcile against.
• Feature store architecture and online-offline parity: a parallel treatment of keeping a fast local store consistent with an authoritative central one.
• ECB digital euro programme pages: primary source for the preparation phase and timeline.
• European Parliament legislative train: digital euro: current state of the legislative file.

By Riju — about