Stablecoins as a payment rail, compared honestly

Sending a stablecoin moves a dollar-denominated balance between two addresses and settles that transfer with finality in seconds or minutes, at a fee set by network conditions rather than by the size of the payment. That is a genuine improvement over some existing methods for some use cases. It is also only one layer of what a payment system does, and the honest comparison requires naming the layers that a blockchain transfer does not provide.

What a payment system does, layer by layer

Authorization confirms the payer is entitled to pay. Clearing establishes what is owed between the parties. Settlement is the actual transfer of value with finality. On top of those sit a set of services that are easy to overlook: dispute resolution and chargebacks, fraud liability allocation, credit provision, refunds, recurring authorization, and compliance including know your customer, anti-money laundering checks, sanctions screening, and the travel rule for originator and beneficiary information.

A stablecoin transfer is very good at exactly one of these. It collapses clearing and settlement into a single event that is final by design, available continuously, and indifferent to which country the recipient is in. Everything else in the list either has to be rebuilt by the businesses using the payment rail or does not exist.

Against card networks

A card payment is an authorization followed by a settlement that arrives days later, and the merchant fee funds a specific bundle: the guarantee that a stolen card does not become the merchant's loss, a dispute process the cardholder can invoke months after the fact, credit extended to the consumer, and a global acceptance network. Interchange is not purely rent; a large part of it buys reversibility and fraud liability.

A stablecoin transfer has none of that. It is irreversible by construction, which is precisely what makes settlement final and also what makes a mistaken or fraudulently induced payment unrecoverable. For merchants, irreversibility removes chargeback losses. For consumers, it removes the protection that chargebacks provide. Both statements are true simultaneously, and which one dominates depends on who is being asked.

Against correspondent banking

A cross-border bank transfer passes through a chain of correspondent banks, each applying its own compliance checks, cut-off times, and fees, with limited visibility into where the payment is and what it will cost on arrival. Value is typically available in one to several business days, and weekends do not count. This is the comparison where stablecoins look strongest, particularly for remittance corridors, business-to-business payments to jurisdictions with slow banking connections, and any payment where a business is waiting on funds it has already earned.

The important qualification is that the honest comparison is not chain fee versus wire fee. It is the total cost of the round trip, including the on-ramp converting local currency into tokens and the off-ramp converting them back at the other end. Those conversions carry a spread, sometimes a fee, and often a compliance process, and in most corridors they dominate the total cost. A near-zero transfer fee between two ramps that each charge a percentage is not a near-zero payment.

The operational realities a business meets

Fees are volatile. A transfer costs what block space costs, so the same payment can be inexpensive one day and materially more expensive during congestion; a layer 2 reduces this substantially and adds its own assumptions about the sequencer and withdrawal path. Payments are address-based and irreversible, so a mistyped or substituted address is a total loss, which is why address poisoning exists as an attack. The issuer can usually freeze an address, which helps in theft cases and means the balance is not unconditionally the holder's. Chain choice is a real decision, because the same nominal token on different chains is different contracts and moving between them involves a bridge with its own history of failures. And a business still has to account for the position, which in most jurisdictions means tracking it as an asset rather than as cash.

Where the model fits, stated without enthusiasm

The pattern that works well is settlement between businesses that already trust each other, across borders, outside banking hours, where reversibility is not required and both ends have solved conversion. Treasury movement between a firm's own entities, payouts to contractors in countries with slow banking access, and settlement between trading venues fit that description. Consumer retail payments fit it least well, because the consumer-protection layer that cards provide is the part a blockchain does not replicate.

Reading payment activity on this site

Settled volume and daily transactions describe how much a chain is carrying, and active addresses describe how many distinct participants were involved, with the caveat that addresses are not people and one entity can control many. Costs are captured by average transaction fee and median transaction fee, and the median is the better guide to an ordinary payment because averages are dragged upward by a few complex transactions. Speed characteristics appear as block time and transactions per second, neither of which equals finality, since finality depends on the consensus rules rather than on how fast blocks arrive.

Two stablecoin-specific figures complete the picture. Stablecoin supply hosted shows how much dollar balance sits on a chain, and market cap to stablecoin supply compares a network's own asset value with the dollar balances it carries. That ratio borrows the shape of an equity multiple and means nothing of the kind: it has no earnings in the denominator, it is not a valuation, and a high or low reading is a description of composition rather than a judgment.

The final article in this track covers the regulatory regimes that govern stablecoin reserves and issuance. See fees for cost data by chain, network activity for throughput, and stablecoins for supply by chain.