There is a difference between a wall that was never built and a door that was built, fitted, and then quietly locked. The first is architecture. The second is a promise. Last month, on a United Nations stage in New York, Charles Hoskinson โ the co-founder of Cardano โ described the proposed digital euro as a possible "financial panopticon," and warned that within a decade Europeans could find themselves unable to buy gasoline because an algorithmic rule deemed the purchase impermissible. The phrase travelled fast through crypto-native media. But the detail that should have travelled with it, and did not, is more interesting: the current draft of the digital euro already excludes programmable spending rules. The switch exists. It is simply off โ for now.
That is the sentence worth auditing. Not the headline, not the quote, not the ten-year prediction. The mechanism, and the conditions under which the mechanism can be re-engaged.
The digital euro is the European Central Bank's project to issue a retail central bank digital currency โ a digital form of cash, denominated in euro, issued directly by the central bank rather than by a commercial bank. It is not a token. It is not a protocol with a governance token and a treasury. It is sovereign money, digitised, and it inherits the trust assumptions of the institution that issues it.
Its status, as of this writing, is legislative draft. Not law. The European Commission's proposal is under negotiation between the European Parliament and the Council, and the parameters that matter most remain unsettled. Two of them define the entire debate. The first is the holding limit โ the maximum amount of digital euro any individual may hold. The second is the merchant rate โ the fee that merchants would pay to accept it. Neither has reached consensus. The pilot is scheduled for 2027, with the earliest possible issuance around 2029.
Beneath the surface, three objectives are pulling against one another. Privacy law โ GDPR โ demands that personal financial data be minimised and protected. Anti-money-laundering rules demand the opposite: visibility, traceability, and reporting. And monetary policy demands that the holding limit be high enough to make the currency useful, yet low enough to prevent a bank run. No single design satisfies all three. Any compromise will disappoint at least one constituency, which is precisely why the parameters remain unagreed and the negotiation drags.
For those of us who have spent years auditing decentralised governance, the architecture here is familiar in the way a photograph is familiar to a map. The digital euro will almost certainly run on a two-tier model: the central bank issues the liability, and commercial banks and payment intermediaries distribute and service it. This is the dominant design across every major CBDC programme, from the Chinese e-CNY to the various sterling experiments. It is efficient. It is also the precise point at which the economic conflicts concentrate, because the intermediary layer is where deposits, fees, and incentives collide.
I spent 200 hours in 2020 mapping the voting centralisation risks inside Compound's governance mechanism. The lesson from that work applies here with uncomfortable precision: the technical specification is never the whole story. The governance of the parameters is.
Let me be precise about what "programmability" means, because the word is doing a great deal of work in this debate. Programmability is the capacity to embed rules into the money itself โ rules about who may spend, on what, when, and in what quantity. A programmable euro could, in principle, refuse a transaction to a sanctioned entity, expire funds if not spent within a period, or restrict a payment to a category of goods.
The draft currently excludes programmable spending restrictions. Hoskinson reads this exclusion as insufficient, and he is technically correct to do so โ not because the exclusion is meaningless, but because of its legal form. An exclusion written into a non-binding draft is a policy preference; an exclusion written into statute is an architectural constraint. The first can be revised by a subsequent draft. The second requires a legislature to reverse itself. His call for a legislative ban is, at its core, a request to convert a political concession into a structural guarantee โ to move the switch out of the room entirely, rather than leaving it on the wall with the light off.
This is why I find the framing of the debate misleading on both sides. The euro's architects point to the exclusion and say the fear is overblown. The euro's critics point to the capability and say the exclusion is temporary. Both are describing the same object from different angles. The capability is real โ a centralised ledger controlled by a single issuer can, by definition, be programmed. The exclusion is real too โ the draft does not currently permit spending restrictions. What neither side will say plainly is that the security of a centralised system is always a function of the restraint of its operator, never of its design. In a decentralised system, the code constrains the operator. In a centralised one, the operator constrains the code. The digital euro is the second kind. Its guarantees are promises, and promises are only as durable as the institution that makes them.
Now consider the holding limit, which is the more immediate economic question and receives far less attention than the surveillance rhetoric. The holding limit exists to prevent disintermediation โ the migration of bank deposits directly onto the central bank's balance sheet, which would starve commercial banks of the funding they lend out. Set the limit low, and the digital euro becomes a curiosity, too constrained to function as everyday money. Set it high, and you invite a slow run on the banking system. This is a genuine dilemma, not a manufactured one, and it is unresolved.
The merchant rate carries the same structure. Set it too low, and commercial banks have no incentive to build acceptance networks. Set it too high, and merchants โ who already resent card fees โ will simply refuse to onboard. The rate is where the intermediary layer's interests are priced. It is also, quietly, where the question of "who gets to transact, and at what cost" is settled. Hoskinson's warning about "asset and transaction discrimination" is usually read as a privacy concern. It is equally an economic one: discrimination is a fee structure by another name.
Compare this to the Chinese e-CNY, which has already run large-scale pilots and settled its design parameters through a single-party process unencumbered by parliamentary negotiation. The euro cannot move that fast, and it should not. But the comparison reveals something the surveillance framing obscures: the digital euro's slowness is its protection. Every month the draft spends in negotiation is a month the exclusion on programmability is debated in public, amended, and potentially hardened. The e-CNY's speed was purchased by removing exactly the friction that now constrains the euro. Where one system traded deliberation for deployment, the other is trading deployment for deliberation โ and for a monetary infrastructure meant to last decades, that is the better trade.
I have written before that most project KYC is theatre, that the compliance burden lands almost entirely on honest users while the determined route around it costs a few wallet holdings. The digital euro raises the same question at sovereign scale. If the system is programmable, the honest user bears the friction; if it is not, the friction is unenforceable anyway. You do not make money safe by making it watchful. You make it watched.
Here is the angle that the crypto community will not enjoy, and which I will state anyway, because the discipline of auditing requires it.
The digital euro will not arrive in 2026. The pilot is 2027; the earliest issuance is 2029. For the next two to three years, its practical effect on any European's financial life is zero. The narrative, however, is already fully priced into the discourse โ and narrative without delivery is a tax on attention. Hoskinson's gasoline scenario is a hypothetical, not a report. It describes a mechanism that could exist, not an event that has occurred. When I audited the Compound governance report, the most useful thing I did was separate the vulnerabilities that were live from the ones that were merely possible. Readers of the digital euro debate deserve the same rigour.
There is a second, more uncomfortable point. The euro's architects excluded programmability in the draft precisely because they anticipated this backlash. That is not evidence of good faith; it is evidence of responsiveness. But responsiveness is itself a constraint โ a legislature that fears its voters is a legislature that will think twice before arming the switch. The panic that treats the exclusion as meaningless removes the incentive for the exclusion to exist.
So the pragmatist's test is this: which is more likely โ that a democratic legislature, bound by GDPR, answerable to 450 million voters, arms a programmable spending restriction by 2036? Or that the fear of that outcome, amplified by figures with commercial interests in decentralised alternatives, drives capital and attention toward privacy tooling that never had a market before? I do not know the answer. But I know which question is being asked, and which is being shouted.
The digital euro is a long negotiation, and the only signal that matters is the one that moves from draft to law. Track the European Parliament's amendments, not the conference circuit. Watch whether the prohibition on programmable spending restrictions is written into binding text, or left as a preference in a superseded document. That single line will tell you more about the future of European money than any speech delivered on any stage.
Hype burns out. The statute remains in the ledger. We audit the logic, because institutions, like code, will always drift toward the path of least resistance โ unless the path is deliberately closed.