Tokenized deposits vs. stablecoins: a false choice

A corporate treasurer wants to pay a supplier on a Friday afternoon. The supplier banks elsewhere, holds part of its working capital in a regulated stablecoin, and prices in dollars. The treasurer's bank has just launched a tokenized-deposit product and would prefer that payment stay inside its own ledger. It will not. The money has to move from a deposit token on one network, through a settlement venue, into a stablecoin held in a wallet the bank does not control. Every leg of that journey has to be screened, recorded, and made defensible to an examiner.

That transaction is the actual product requirement. The debate banks are having instead, tokenized deposits or stablecoins, is the wrong frame for it.

Two instruments, two jobs

The instruments are not interchangeable, and the differences matter to a risk team.

A tokenized deposit is a claim on a specific bank, recorded on a ledger. It sits on the bank's balance sheet and inside the supervisory framework that already governs deposits. It is subject to the bank's own BSA program, and it never leaves the regulated perimeter the bank already defends. For a large bank, that is the entire point: on-chain money that does not concede the deposit franchise to anyone.

A stablecoin is a bearer-style instrument backed by reserves, transferable to anyone with a wallet, governed under a payments-and-reserves regime: the GENIUS Act framework in the US, MiCA's e-money token rules in the EU. It is portable in a way a deposit token is not. That portability is its value to the holder and its risk to the issuer, because the instrument moves into addresses no one underwrote.

These map to different needs. A tokenized deposit settles between known, banked counterparties who want finality without surrendering the deposit relationship. A stablecoin buys reach: payment to a counterparty the bank has no relationship with, often in a market the bank does not operate in. Any institution of size will need both over a normal week.

The big banks have already chosen, and named the gap themselves

Our reading of the US landscape is that the largest banks are not neutral here. Tokenized deposits are the preferred instrument, and the preference is explicitly defensive: a deposit token keeps balances inside the regulated perimeter, where a stablecoin would let them migrate out. The coordination is visible in the architecture. Tokenized-deposit work tends to route through shared market infrastructure rather than two dozen private chains, and the same names recur on the same underlying ledger stack: Kinexys on one side, GS DAP on the other, both built on Canton and Digital Asset. Goldman has signaled it intends to spin its platform out to be industry-owned rather than keep it captive.

So the instrument question, at the top of the market, looks largely settled in favor of deposits. What is not settled, and what nearly every institution we have studied names as the missing piece, is connectivity. Banks are building islands of on-chain money. Each island works. The deposit token clears cleanly inside its own network. The problem is the water between them: moving value from a closed bank network to an open one, from a deposit token to a stablecoin, from one consortium's ledger to a counterparty on a different chain entirely. The hard engineering and the hard compliance both live in the crossings, not on the islands.

This is why "pick a winner" is the wrong instinct. A bank that bets everything on its tokenized-deposit network still has customers who hold stablecoins, suppliers who demand them, and counterparties on chains it does not run. Supporting one instrument does not exempt the bank from interacting with the others. It only determines whether the bank does so deliberately or reactively.

Interoperability is a compliance problem before it is a technical one

The reason crossings are hard is not message formats. It is that every form of on-chain money carries its own obligations, and moving between them does not reset the clock. It stacks the requirements.

Return to the supplier payment. When deposit-token value becomes stablecoin value bound for an external wallet, the bank inherits a chain of obligations. OFAC screening has to cover the named beneficiary, the destination address, and that address's transaction-graph exposure, direct and indirect, because name-matching alone misses a wallet that sits two hops from a sanctioned mixer. The Travel Rule applies once the transfer clears its threshold (FATF's roughly $1,000 floor, the US BSA's $3,000), with originator and beneficiary information transmitted to the receiving institution, which may be a stablecoin issuer operating under a different regime than the one that governed the deposit leg. KYT monitoring has to read on-chain behavior the bank's core systems were never built to see. The whole sequence has to produce an immutable, exportable, examiner-ready record, because under the current third-party-risk posture the bank must defend both the transaction and the vendor to a supervisor who wants evidence, not explanations.

None of that gets easier because the bank standardized on tokenized deposits. The deposit leg is clean precisely because it stays inside the perimeter. The compliance burden materializes at the boundary, the moment value crosses into an instrument or a network governed differently. The institutions that "won" the instrument debate by going all-in on deposits have not avoided the hard part. They have relocated it to the edge of their own network, where it is easiest to overlook.

The design goal, then, is not a universal token. It is a layer that lets a bank hold value in whatever form a given transaction requires while keeping one consistent control surface across all of them: the same screening, the same KYT, the same Travel Rule handling, the same audit trail, whether the leg is a deposit token, a stablecoin, or a settlement on a network the bank chose to connect to but does not own. The bank stays the principal. The instrument becomes a detail of the trade rather than a re-architecture of the compliance program.

What this asks of strategy and risk teams

A few things follow.

Treat instrument selection as portfolio composition, not a binary. The question is which forms of on-chain money the bank supports, in what order, for which corridors, and the answer will shift as customers and counterparties move. Architect for plurality so that adding the second instrument is a configuration change, not a program.

Locate the compliance investment at the boundaries. Controls that live inside a single network are table stakes. The obligations that will actually be examined are the cross-instrument, cross-network, cross-regime ones. Build screening, monitoring, and recordkeeping to span the crossings, integrated with the analytics providers examiners already expect to see.

Keep the connecting layer neutral with respect to the instrument. Anything that forces the bank to favor one form of money, or onto someone else's chain to obtain interoperability, reintroduces the lock-in the bank was trying to avoid. The bank should be able to decide where it settles and in which asset, on a layer that extends what it already runs.

This is the worldview we build from at Suave. We are early, and we say so plainly. Finance is moving on-chain in fragments, and the durable value is in connecting the fragments without asking the institution to give up control of any of them. The bank keeps its networks, its keys, its stack, its customers. The instrument it uses on a given Friday afternoon is its call.

The false choice is believing the choice is final. The institutions that age well will be the ones that stop asking which form of on-chain money wins and start asking how value moves cleanly between all of them, inside the perimeter they already answer for.