A CBDC is the most consequential piece of financial infrastructure most people will never choose to use. No one will download it, no one will opt in, and in most designs it will arrive as an option inside an app you already have. That is precisely why the design choices being made now matter more than the debate about whether it happens.
Three design questions decide almost everything
Wholesale or retail. Wholesale CBDCs settle between banks and institutions, replacing the slow parts of interbank settlement. They are unexciting and largely uncontroversial. Retail CBDCs would be held by people, competing directly with bank deposits and with cash. Almost all the political heat is here.
Account-based or token-based. An account model means the central bank (or its intermediaries) keeps a balance per person and knows every transaction, like a bank account. A token model tries to look more like cash: the instrument carries the value and does not require a central ledger of identity-linked balances. Most deployed designs are account-based, and the privacy difference between the two is the whole argument.
Direct or two-tier. Direct means citizens hold claims on the central bank. Two-tier keeps commercial banks and payment providers as the interface, with the central bank as the back end. Nearly every serious project has chosen two-tier, because direct central bank accounts would drain deposits from the banking system during stress — and that is a stability problem, not a technical one.
What a CBDC is not
It is not cryptocurrency. There is no consensus mechanism to trust, because there is no question of who is right: the issuer is the authority. It is not a stablecoin either, though they compete for the same payments. A stablecoin is a private claim on assets held by a company; a CBDC is a sovereign claim with monetary policy attached.
Where the two overlap is the interesting part. If a central bank offers a programmable domestic payment rail for free, a stablecoin's use case narrows to what the state cannot or will not serve: cross-border transfers, dollar access in countries with unstable currencies, and programmable settlement that no single authority can freeze.
The privacy question is not binary
Cash is anonymous by construction. Card and app payments are recorded by intermediaries. A CBDC starts from a choice, and the choice is about who can see what: the central bank only in aggregate, the intermediary at the individual level, or third parties under legal process.
Some designs are actively exploring privacy-preserving options — limits on holdings to keep the instrument from becoming a savings vehicle, offline payment modes, and commitments not to attach programmable restrictions to individual payments. Those commitments are policy, not code, which is why the loudest critics and the most enthusiastic designers are fighting about governance rather than cryptography.
Where the real change is
Domestic retail CBDCs mostly reproduce payment systems that already work. The case where a CBDC genuinely improves something is cross-border: correspondent banking is slow, opaque, and expensive, and multi-CBDC settlement platforms aim at exactly that.
For anyone building in crypto, the reasonable posture is neither dismissal nor doom. Assume coexistence: sovereign digital cash for domestic retail, private and crypto rails for the parts of the world where a domestic rail is not what people need.
What to watch
The design documents, not the headlines. A pilot announcement tells you a testing programme exists. What matters is whether the build is two-tier, whether holdings are capped, whether privacy guarantees are statutory, and what happens during a bank run.
Next: the stablecoin side of the same question. Join our Discord.
References
Next, read Crypto Insurance 101. More in Web3 Economics.