The foreign exchange market trades roughly $7.5 trillion per day, making it the largest financial market in the world by daily volume — larger than all global equity markets combined, larger than bond markets, larger than derivatives by most measures. It operates continuously, around the clock, across every currency pair imaginable, and it is almost entirely unregulated in the traditional sense. There is no central exchange, no consolidated tape, no regulator reviewing every transaction. It is a dealer market, which means that prices are set by a small number of large banks that make markets between themselves and then distribute those prices, with a markup, to everyone else.
The markup is the point. The difference between the price at which a dealer buys a currency and the price at which it sells — the bid-ask spread — is the primary mechanism by which foreign exchange intermediaries extract value from the market. For the most liquid pairs, like EUR/USD, the interbank spread is measured in fractions of a pip — a tenth of a basis point or less. For the currencies that most of the world's population actually uses — the Nigerian naira, the Philippine peso, the Kenyan shilling, the Bangladeshi taka — the spread can run to several percentage points. A Filipino worker sending remittances home from the United States might lose five to eight percent of the transfer to exchange rate margins and fees before the money arrives. That is not a rounding error. It is a structural tax on labour mobility.
Stablecoins don't solve every part of this problem. But they address the part that the dealer banks have spent decades insisting is technically unavoidable: the cost of moving value across currency boundaries. That cost, it turns out, was primarily the cost of the intermediaries involved, not the cost of the transfer itself.
How the FX Market Actually Works
The foreign exchange market has a tiered structure that most participants never see. At the top sits the interbank market — the layer at which the largest financial institutions trade directly with each other, accessing the tightest spreads available. Below that sits the dealer-to-client layer, where banks and non-bank financial institutions quote prices to corporate treasuries, institutional investors, and hedge funds. Below that sits the retail layer, where everyone from importers to tourists to remittance senders gets the widest spreads and the least transparency.
The critical feature of this structure is that price formation happens at the top and deteriorates at every subsequent layer. A corporate treasury executing a $50 million EUR/USD trade gets a price close to the interbank rate. A small business wiring payment to a European supplier gets a price that may be fifty to a hundred basis points wider. A retail customer at a bank branch gets a price that may be two hundred to three hundred basis points wider still, plus a fixed transaction fee. The spread at each layer reflects the dealer's cost of execution, its inventory risk, and — significantly — its ability to price discriminate based on what each counterparty will accept.
Price discrimination works in FX because transparency is structurally absent. A corporate treasury with an FX desk has Bloomberg terminals, multiple bank relationships, and the ability to request competing quotes. A retail customer has the rate posted on the bank's website and the option to not transact. The dealer knows that the corporate client will comparison-shop; the dealer also knows that the retail customer probably won't. The spread reflects that knowledge.
The Correspondent Banking Layer
Cross-border payments add a second intermediation layer on top of FX conversion. To move money from one country to another, a bank that lacks a direct relationship with the receiving bank must route the payment through one or more correspondent banks — institutions that maintain accounts in both jurisdictions and charge fees for the service. A payment from the United States to the Philippines might route through a US correspondent, a regional correspondent in Southeast Asia, and then the local bank, with each hop extracting a fee and introducing settlement delay.
SWIFT, the messaging network that most international bank transfers use, does not itself move money — it sends instructions. The actual movement of funds happens through a chain of debit and credit entries in correspondent accounts, a process that takes one to five business days depending on the corridor and the banks involved. During that period, the funds are effectively in transit, earning nothing for the sender and generating float income for the correspondent banks holding them.
The World Bank estimates that the global average cost of sending a $200 remittance was 6.2 percent as of early 2024. The G20 set a target of reducing that average to three percent by 2030. That target will almost certainly not be met through reform of the correspondent banking system, which has had decades of G20 targets and sustained pressure to reduce costs and has responded with incremental improvement at best. The correspondent banking model is structurally resistant to compression because its costs are the fees of multiple intermediaries, each of which has independent interests in maintaining its margin.
The foreign exchange and correspondent banking systems are not expensive because currency conversion is technically difficult. They are expensive because a small number of institutions have controlled the infrastructure through which it must be done, and that control has allowed them to price the service at whatever the market will bear — which, for most participants, is quite a lot.
What Stablecoins Change About This Picture
A stablecoin transfer bypasses the correspondent banking chain entirely. When a worker in the United States sends USDC to a recipient in the Philippines, the transaction settles on the blockchain in seconds, with a gas fee measured in cents rather than percentage points. The recipient can then convert USDC to Philippine pesos through a local exchange or, increasingly, spend directly from a stablecoin wallet at merchants that accept them. The correspondent banks, the SWIFT messaging fees, and the settlement float all disappear from the equation.
The FX conversion step — converting dollars to pesos — still exists, but it now happens in a different market. Decentralized exchanges (DEXs) like Uniswap operate automated market makers (AMMs) that provide continuous liquidity for token pairs, with spreads that are visible, auditable, and competitive in a way that bilateral dealer quotes are not. Centralized cryptocurrency exchanges like Coinbase and Kraken provide FX-adjacent services with fee structures that are published and comparable. Neither is a perfect substitute for a bank's FX desk for large institutional transactions. For the $200 remittance use case, they are already superior on price in many corridors.
The evidence is not purely theoretical. Bitso, the Mexican crypto exchange, processes a substantial portion of USD-MXN remittance volume between the United States and Mexico using stablecoin rails. The company's reported effective rate — the total cost to the sender including exchange rate margin and fees — is materially below the traditional bank and money transfer operator alternatives in that corridor. Strike, the Bitcoin-based payments application, has offered sub-one-percent remittance costs in corridors where its liquidity is sufficient. These are not proofs of concept. They are functioning products moving real money for real users at costs that the incumbent system cannot match.
The Institutional FX Question
The remittance corridor is the obvious near-term application, but the structural question extends to institutional FX as well. Corporate treasuries execute hundreds of billions of dollars in currency conversions annually to manage exposure arising from international operations. Each of those transactions involves a dealer spread, a settlement delay, and a counterparty relationship cost. For large multinational companies, FX costs represent a material drag on international operations — one that treasury departments work to minimize but can never fully eliminate within the current market structure.
Tokenized currencies — digital representations of major currencies issued on public or permissioned blockchains — offer a different settlement model. If a US company's euro revenues and dollar operating costs can both be held as blockchain tokens, conversion between them becomes a programmatic token swap rather than a bilateral dealer transaction. The settlement is atomic, the rate is determined by a transparent market mechanism, and the counterparty risk associated with two-day FX settlement is eliminated.
JPMorgan's Onyx platform has been processing intraday repo and FX transactions on a permissioned blockchain since 2022, with participation from institutional clients including Goldman Sachs and BlackRock. The platform's reported transaction volume had exceeded $700 billion by late 2023. This is not stablecoin infrastructure in the consumer sense — it is bank-issued digital money on a permissioned ledger. But it demonstrates that the institutional appetite for programmable settlement in FX and money markets is real, and that major banks are investing in the capability rather than dismissing it.
Where the Dealer Model Survives
The FX dealer's position is not uniformly threatened. For the most liquid major currency pairs and the largest transaction sizes, the interbank market already provides extremely tight spreads, and the operational infrastructure of major FX desks — real-time risk management, regulatory compliance, credit intermediation — provides genuine value that on-chain alternatives cannot yet fully replicate. A central bank executing a billion-dollar reserve rebalancing needs counterparties with deep balance sheets and established settlement infrastructure, not a DEX with liquidity pools that may or may not be adequate for a transaction of that size.
The dealer model is most vulnerable at the edges: the long tail of less liquid currency pairs, the retail and small-business segment where price discrimination is most aggressive, and the corridor-based remittance business where the cost of the existing infrastructure is most visible to the people paying it. These are not peripheral segments. They represent the majority of the world's population, most of the world's cross-border labour income, and a substantial portion of global trade finance for small and medium enterprises.
The dynamic is the same one that has played out in every financial market that blockchain infrastructure has reached. The incumbents retain their position in the high-complexity, high-margin institutional segment longest. The retail and mid-market segments compress first and fastest, because that is where the alternative is most obviously better and the switching cost is lowest. By the time the incumbent recognizes the threat as existential rather than marginal, the customer base that supported the margin structure has already migrated.
The FX market has operated on its current terms for decades in part because there was no plausible alternative for anyone who needed to move money across currency boundaries. That condition no longer holds unconditionally. The alternative exists, it works, and it is getting cheaper and more liquid every year. The bid-ask spread that has funded dealer FX operations for a generation is now a competitive target, not a fact of nature.