Digital euro from the vendor perspective, revisited

Four months on: what changed, what I got wrong, and what I would tell a team today. Published 20 November 2026 · Digital euro

First published on LinkedIn. It revisits my July 2026 article on the same topic.

Digital euro symbol at the centre of a circuit-style map of Europe, with a phone paying at a card terminal

When I wrote my previous article in July, I wanted to move away from the politics of the digital euro and look at the build. Who has to do what, which systems need to change, and where the real implementation risk sits.

A few months later, there is more to work with. Parliament has confirmed its negotiating mandate, the pilot has named participants and two network service provider groups have been selected. We also have more technical documentation to read, question and eventually turn into delivery plans.

It is a living thing, it changes from month to month, but my scepticism has not disappeared. Neither has my view that organisations should prepare. If this is going to be done, it needs to be done the right way. Those positions fit together quite comfortably. As a consultant, I can question the economics and the scope while helping a client understand what it may need to deliver. In fact, I think that is exactly the job.

So, time to revisit the build.

Where things stand today

The ECB continues to prepare for a 12-month pilot starting in the second half of 2027 and aims to be ready for potential first issuance in 2029. That timetable remains conditional on the legislation, and issuance still requires a separate decision. In its November statement to Parliament, the ECB was still calling for the legislative package to be concluded.

For a delivery team, those distinctions matter. Pilot readiness, production readiness and a decision to issue are different milestones. A programme plan that turns them into one date will look simple until someone has to deliver against it.

I would prepare against the published direction, but I would not sign away the uncertainty.

The connectivity layer now has names

On 30 October, the ECB announced two selected network service provider groups:

  • Nexi Group Payments and Deutsche Telekom Global Business Solutions.
  • Telefónica Global Solutions and Indra Soluciones Tecnologías de la Información.

Their scope includes connecting participating PSPs and national central banks to the Digital Euro Service Platform, the DESP, and supporting communication between its components. PSPs would have the option to connect through either provider.

It is an important milestone, and it also makes one part of the vendor discussion more concrete.

These are the two selected network provider groups, not the entire digital euro supplier ecosystem. Other suppliers had already been selected for components including alias lookup, fraud and risk management, the app and SDK, offline functionality and secure payment-information exchange.

For organisations planning participation, connectivity now belongs in the architecture and procurement workstream. Who contracts for it? Who operates it? How does it fit into the existing security model? What happens when connectivity is available but an application service is not? Who owns the incident across the boundary?

And please, do not confuse a connectivity contract with a working PSP proposition. The wallet, ledger integration, funding flows, merchant acceptance and operational support still need owners.

Decide what business you are entering

I would still start with the same question: are you distributing, acquiring, or doing both?

But there is a further distinction worth making. The current draft does not put every PSP under the same obligation. It differentiates mandatory distribution by specified credit institutions, funding and defunding support, and voluntary participation by other eligible PSPs.

A bank preparing for a distribution obligation has one starting point. A PI or EMI considering an optional product has another. A merchant technology provider deciding whether to build acceptance capability has a third.

Before anyone produces an estimate, agree which problem the organisation is solving. Compliance readiness? Customer retention? A new acceptance product? A managed service for smaller institutions?

Ignore this, and you will end up with a couple of separate silos. Trust me.

The funding flows still deserve attention

In the previous article, I spent time on the waterfall mechanics. I still think this is where a lot of the practical difficulty sits.

First, a terminology correction. Reverse waterfall funds a payment when the payer does not have enough digital euros. Waterfall converts digital euro receipts into commercial bank money, for example when incoming funds exceed an applicable holding limit. I used “reverse waterfall” too broadly on the merchant side previously. The distinction matters because the flows, dependencies and failure cases differ.

The holding limit also remains a moving parameter. €3,000 has been used in hypothetical analysis. It is not a final, agreed production maximum. My advice to a team would be straightforward: make the limit configurable and keep the funding logic separate from assumptions about its eventual value.

Then test the awkward cases. That is where I would spend workshop time. A happy-path diagram is useful, but it does not tell me whether your operation can recover.

The merchant sweep changes the work, not the need for it

Parliament’s negotiating position allows businesses to accumulate incoming digital euros temporarily, for up to 24 hours, to enable batch defunding. That is different from assuming every merchant payment must trigger its own immediate conversion.

Batching may reduce conversion frequency. It also creates another set of operational questions. Which receipts are included in which sweep? How are they reconciled? What happens when defunding fails? How are refunds funded? Who monitors funds approaching the permitted holding period?

I would want those answers before accepting an estimate that treats merchant settlement as a small addition to the payment flow. Anyone in the card business today who is responsible for reconciliation and settlement knows what a pain this can be, even without being mandated to move money under time pressure.

Merchants need understandable reporting and support, and acquirers need an operating model. Someone has to connect the two.

Read the documents, including the disclaimers

I previously advised teams to read the rulebook and API specifications rather than rely on press releases. I stand by that, with a qualification.

Draft rulebook v0.91 explicitly says it should not be used for pilot or implementation purposes. Dedicated pilot documentation applies, and the eventual scheme remains dependent on the legislation.

So document status belongs in the delivery process. A requirement from a legislative negotiating position is different from a pilot interface requirement. A draft scheme assumption is different from a final contractual obligation.

Take the account model, for example. The current draft assumes one account per individual, but the rulebook itself flags the difference from the multiple-account possibility in the Commission proposal. That is manageable if the assumption is visible. It becomes expensive if it turns into changes to database design, customer journeys and fixed-price commitments.

Every significant assumption should have an owner, an impact assessment and a trigger for revisiting it. Basic delivery discipline, applied consistently.

Offline remains a separate workstream

Oh boy. Offline still deserves its own scope, security assessment and estimate. If there is one critical part in all this, this is it.

Locally stored value, secure hardware, device-to-device transfers and funding or defunding create a different design problem from an online account-based payment. Offline privacy also means teams cannot simply assume the same transaction visibility and monitoring model.

The sensible planning position is to treat offline holdings as cash-like exposure, without promising online-style recovery. Recovery, replacement and compensation need to be checked against the actual technical design and applicable liability arrangements.

The customer explanation will matter too. I do not like it too much, and I work in this field. Neither will customers, I assume. If two modes sit inside the same wallet but behave differently when something goes wrong, support teams need more than a paragraph in the terms and conditions.

Reuse needs to be demonstrated

There is useful work around reusing European standards. The ECB has agreements involving the European Card Payment Cooperation, nexo standards and the Berlin Group. That can help acceptance and integration.

I would still challenge any estimate that jumps from “existing standard” to “small change.” PwC’s March 2026 cost-study follow-up makes a relevant point: banks already assumed reuse and outsourcing. New implementation work, integration, certification and governance do not disappear because some underlying standards are familiar.

Show me the reusable component, which interfaces remain unchanged, the certification evidence and the responsibilities retained by the PSP. Then we can discuss the saving.

The same applies to managed services. Buying a service can be the right decision, but the contract still needs to explain what the provider delivers, what remains with the client and how changes are priced.

The cost argument is still open

I remain unconvinced that the lower estimates tell an organisation enough about its own exposure.

The ECB estimates banking-sector investment at €4 billion to €5.8 billion. The banking-association-commissioned PwC study estimates around €18 billion in bank change costs, an order of magnitude confirmed in its March 2026 follow-up. These figures reflect different assumptions and should not be treated as interchangeable measurements of the complete ecosystem. By all means, no.

I would carry forward the “north of €60 billion” reference from my previous article even without a clearly identified source and scope. My underlying concern remains.

Who has counted merchant migration? Internal staff? Parallel operation? Certification? Ongoing provider charges? Changes after the baseline moves? Losses during the process, because other projects must stop? Or, if not, the additional resources that must be engaged?

The customer still needs a reason

My empty-wallet concern remains.

Mandatory availability does not establish customer demand. Broad acceptance does not tell us how frequently people will choose to use the product. This is the KEY issue.

What benefits are there for the end customer? Lower prices due to lower costs for the merchant? I have yet to see that scenario. I never saw it in my 20 years in payments and banking, not once. We all remember how the interchange cap sent retail prices plummeting in the EU. Not.

Meanwhile, instant payments are already part of the euro-area baseline. The principal bank deadlines applied in 2025, while euro-area PIs and EMIs have a separate April 2027 deadline. The commercial question therefore gets harder. And everyone is using it. It is soon to become the default rail in every banking app (we still need to toggle that annoying “use instant” switch or dropdown).

What does the customer gain? What does the merchant gain? What happens to the business case if uptake is slow, balances stay low, and users rely heavily on automatic funding? I would expect a serious programme to model those cases, rather than hide them behind an adoption assumption.

What I would tell a team today

Start with your role and exposure. Build the operating model before pretending the estimate is complete.

Separate online, offline, acceptance, connectivity and funding workstreams. Identify the ledger and reconciliation boundaries. Make vendors demonstrate reuse. Design the failure cases early.

There is preparatory work worth doing now. There are also dependencies that should remain explicit until the legal, technical and commercial baseline is firm(er). The draft status of the scheme is a reason for controlled investment, not inactivity.

My view remains that this can be built. Whether the final design produces enough value for its cost is still a legitimate question. Organisations preparing for it should keep asking that question while doing the engineering properly.

The teams I would trust are the ones that can explain what they know, what they are assuming, what they are buying and which risks they still own. And what they do not know.

Because, in the end, as the famous phrase says: “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.”


About the author

Goran Bosankić, Founder of PayconGoran Bosankić, ETA CPP, is the founder of Paycon, an independent payments and banking advisory. Fifteen-plus years across core banking, digital banking, BaaS, card acquiring, payment orchestration, and on-chain rails, on both sides of the stack. Previously CEO of a regional banking-technology group and CRO of a payment orchestration platform. LinkedIn · [email protected] · More about Paycon


If you are scoping a digital euro build as a distributing PSP, an acquirer or a vendor, and want an independent read on the rulebook, the DESP integration or the offline piece, get in touch. See also Paycon’s digital euro advisory and stablecoin services, the companion piece on fast payments at the European point of sale, the analysis of deposit tokens and stablecoins, and the rest of the Insights archive.