On a Saturday in March, oil prices spiked 30% after the Iran conflict escalated. Traditional exchanges were closed. FedWire was closed. CLS was closed. Traders who needed to act didn't wait. They migrated to a DeFi platform and processed $1 billion in weekend oil futures volume over two days. JPMorgan's own analysts flagged the migration.
That should unsettle every large bank. It should also focus them.
The shift
For the better part of three years, most bank strategy documents treated stablecoins as a threat — deposit flight risk, regulatory uncertainty, disintermediation of the core balance sheet. The posture was defensive. The strategy was wait and see.
That posture is now visibly costing money.
In May 2026, SoFi Bank — a nationally chartered, OCC-regulated institution — launched SoFiUSD to its nearly 15 million members. Reserves held 1:1 at the Federal Reserve. Live on Ethereum and Solana. Available through the standard mobile app, not buried in a crypto tab nobody visits. The stablecoin launched alongside tokenized deposits, with customers able to swap between the two 24/7 and redeem either at par.
At roughly the same time, Coastal Community Bank began replacing part of its correspondent banking infrastructure with stablecoin rails. Customers initiate cross-border payments, the sending bank runs standard KYC/AML screening, and settlement happens via stablecoin wallet transfer instead of through Nostro accounts pre-funded with idle capital at a GSIB — around the clock, every day of the week, at a fraction of the cost.
Two different institutions. Two different strategies. The same underlying conclusion: stablecoins are a revenue line, not a liability to manage.
Two kinds of money
The clearest framing for why both belong inside a bank is deceptively simple: tokenized deposits are money that rests. Stablecoins are money that moves.
Tokenized deposits stay inside the institution that issued them. They earn yield, carry clear protections, and handle unlimited volume — but they cannot leave the bank's walls. Stablecoins are open-loop: they exist on any compatible wallet, in any country, at any hour. These are not competing products fighting for the same use case. They're complementary rails serving structurally different ones.
Banks that understand this distinction are already monetizing both sides. JPMorgan Kinexys has processed over $3 trillion cumulatively since 2020. Standard Chartered offers custody, off-ramping, and stablecoin card infrastructure. SocGen launched MiCA-compliant EUR and USD stablecoins. Deutsche Bank clients are asking for stablecoin interoperability. Deposit flight — the fear that drove three years of avoidance — hasn't materialized. Coinbase has offered 4%+ yield on USDC for years. The customers who wanted yield left long ago. The ones who stayed, stayed for the relationship.
Three business cases, all active
The stablecoin opportunity for banks has moved from theoretical to operational across three distinct models.
Local payments access. Stablecoin on/off-ramping and card infrastructure generates direct fee revenue from servicing the existing demand of crypto-native businesses. For regional and community banks with crypto clients, this is the simplest entry point and the fastest path to accretive revenue.
Correspondent banking replacement. Non-GSIBs currently pay $5–$50 per transaction through their correspondent, pre-fund Nostro accounts with capital sitting idle in foreign currencies, and absorb investigation fees every time automated filters generate a false positive on a legitimate transaction. Stablecoin wallets eliminate all three friction points at once. The settlement message still travels the same path — KYC/AML checks happen as before — but the actual value movement is bilateral, 24/7, and at a fraction of the cost. Coastal Community Bank's live model is the proof case.
Off-hours settlement. Fiat doesn't move on weekends. Prices do. The oil weekend is not an edge case — it's a preview of what systematic price dislocation looks like when traditional rails are closed and stablecoin rails are open. Weekend FX typically commands 5–20 basis points. A bank-issued stablecoin scoped to institutional clients and off-hours windows, backed by wholesale deposits, could generate real fee revenue from a window that currently generates nothing. It would also keep that volume inside the institution rather than routing to DeFi.
The gap that remains
One of the most thorough recent analyses of the bank stablecoin opportunity ends with a candid admission: compliance is the obvious gap the author didn't have space to address. Banks adding stablecoin capabilities will need wallet screening, chain analytics, and travel rule tooling alongside their existing AML stack. The tools exist. The operational lift is real. Any honest business case includes it.
But the gap is more structural than that framing suggests.
Wallet screening and chain analytics are post-transaction instruments applied to rails that are, by design, irreversible. Blockchain finality is not a quirk of implementation. It is a structural property of the infrastructure. There is no central administrator who can execute a forced balance clawback from a settled transaction on a public ledger.
And yet Regulation E and Regulation Z still apply. Consumer dispute rights exist regardless of what rails a payment ran on. A bank that issues a stablecoin card or enables stablecoin payments inherits the dispute obligations of the legacy system while operating on infrastructure that doesn't support the legacy resolution mechanism.
The current industry workaround — decoupling card authorization from on-chain settlement, absorbing disputes on the issuer's balance sheet, time-buffering before finality — functions at low volume. It does not scale. And it breaks entirely in the context of agentic commerce, where autonomous systems are initiating the transactions. Post-transaction chargeback logic assumes a human customer who noticed something wrong and filed a claim. It was designed for a world where the paying party is a person making a conscious decision.
What the infrastructure requires
The compliance architecture for stablecoin banking is not the existing stack plus wallet screening. It requires a structurally different approach: one where the trust conditions governing a payment are enforced before the transaction fires, not reviewed after it settles on an immutable ledger.
That means authorization logic that reflects the actual structure of the relationship — who authorized the transaction, under what conditions, against what counterparty commitments, within what spending limits. The dispute mechanism then becomes a function of the authorization architecture, not a manual process applied after irreversible settlement has already occurred.
The business case for stablecoins in banking is real. The regulatory environment is moving — the GENIUS Act has passed, the CLARITY Act is advancing. The revenue models are live, not theoretical. Accenture estimates that up to $13 trillion in transaction value could shift to alternative payment methods by 2030, with $13 billion in payment fees at stake for institutions that don't adapt.
What is not yet ready is the layer that makes programmable money trustworthy at institutional scale. The banks that build that infrastructure alongside the revenue model will have a durable operational advantage. The ones that treat compliance as an afterthought will find themselves with the business case but not the ability to operate it.