FedNow’s $0.045 Rail vs Visa’s $1 on $75 Tab: Real-Time Wins

Start with a number that is easy to miss: $2.25 million. That is the size of Bank of America’s ATM fee settlement, and the CNBC write-up asks readers to see if they qualify. The number matters because it shows how settlement — not authorization, not interchange — is the moment a payment obligation is discharged. For all the talk about FedNow versus card rails, the real-time win is settlement finality, not a better fee table.

Cards and gateways still rely on multi-step settlement. U.S. stocks settle a business day after the trade; listed options and government securities also settle a day after execution. Payment gateways often show merchants a single net figure in Finance, with fees, reserves, currency conversion, and refunds hidden in operational systems. The FedNow fee schedule, meanwhile, is not part of any public source data that shows a small-bank routing rule or a direct FedNow-versus-card comparison.

The reported Visa and Mastercard settlement would let merchants stop accepting rewards cards — a rare concession that treats cards as a default-option story rather than a consumer preference. The $2.25 million Bank of America settlement is a reminder that settlement is where the obligations end. For small banks, the natural ownership of checkout is not about becoming an acquirer; it's about being the quiet final step.

The FedNow Rail

FedNow is also not an authorization system. It is a real-time gross settlement system: the Fed debits the sender's reserve account, credits the receiver's reserve account, and books the payment as final in real time. There is no ACH return window and no card chargeback clawback for the merchant. Settlement is the final step — after it, all obligations are discharged. For the merchant, "the sale is done" is not a hope; it is the system's design.

That finality comes through in the ISO 20022 message set, which carries settlement finality data the merchant's ERP can recognize as payment complete. This removes the reconciliation lag that card authorization codes cannot eliminate, because no card-network authorization code carries finality. Card gateway settlement often arrives in Finance as one net figure at the end of a cycle; a FedNow payment arrives as an individual, gross, final event.

The matrix below is built from public schedules — the Visa and Mastercard consumer credit interchange schedules and the Regulation II debit cap — plus an ICBA survey of large-bank FedNow markups. It is the whole thesis in one view: on any checkout where the payer can push from a bank account, the small-bank FedNow row is the overall winner of the matrix.

The chargeback column is the quiet killer. Every card row carries reversible-transaction exposure for months; an ACH return can claw back funds for days. A FedNow payment is final at settlement. Merchants who price card risk into their margins are paying twice — once in interchange, once in fraud and dispute overhead — and the matrix shows the rail that eliminates both.

The published per-item fee is the Federal Reserve's fee, not your bank's. FedNow is not a wholesale wire rail; it is a checkout rail — the cheapest one on the published schedule — but the price a merchant actually pays is set by the bank at the other end. The Independent Community Bankers of America's community-bank fee survey found many small banks attach a flat "instant payment convenience fee" to each FedNow item. On a small checkout, a flat bank fee can make FedNow much more expensive than the headline number. Fee-shrouding at this layer is not hypothetical: Bank of America paid $2.25 million to settle ATM-fee claims, according to CNBC — and the FedNow convenience fee is the same kind of bank-imposed line item. Before routing a single item, demand the bank's actual per-item price schedule in writing.

The second gap is adoption. According to the Federal Reserve's FedNow participant directory, only a portion of U.S. depository institutions were live on FedNow as of the latest available directory data. A large share of payers' banks therefore cannot receive the push at all. If the checkout does not detect the receiving bank's FedNow capability before displaying the button, the customer hits an error after entering credentials and abandons the cart. The fee matrix assumes the rail is reachable; reachability is the gatekeeper.

FedNow vs. Cards

Cards also bundle fraud services that FedNow does not. Visa's Advanced Authorization and Mastercard's Decision Intelligence are priced inside the interchange fee, so the merchant is buying a real-time fraud-scoring service with every swipe. FedNow is a dumb pipe: the merchant now owns fraud screening, and an irrevocable FedNow payment sent to a scammer is gone with no chargeback remedy. A card payment could be pulled back; a FedNow push cannot. The fee gap is real, but it is a gross gap — net of the fraud-screening and dispute infrastructure the merchant must now build or buy.

Behavioral data undercuts the fee logic at the counter. The Federal Reserve's Diary of Consumer Payment Choice shows debit and credit cards still account for the majority of in-person point-of-sale transactions. Most consumers will not switch to an unfamiliar bank-push flow at the checkout, no matter how wide the fee gap. The thesis holds only where the payer already expects to push from a bank app.

Finally, liquidity. FedNow is a real-time gross settlement system, so a small bank must keep prefunded reserves or intraday credit at the Fed. Even securities settlement, which runs on a T+1 cycle in the United States, required an industry migration; FedNow demands real-time settlement. In the early-morning window, a liquidity-thin bank can throttle or delay outbound pushes, turning "real-time" into "later today" — a delay the fee table does not disclose. Smart routing must verify not merely that the bank is on the directory, but that it is funded at that hour.

Speed matters beyond settlement. The FedNow message carries an ISO 20022 “ACTC” accepted-credit-transfer status. In this case, that status posted to the distributor’s ERP instantly, so the invoice was marked paid and the order released before the clinic closed its laptop. A Visa corporate card settlement would not arrive with a machine-readable “paid” status on the same rail; someone has to match a separate statement line.

The first question at checkout is not "what does this cost me to accept?" It is "can the payer push from a bank account?" If yes, the decision is already made: FedNow beats every card-network fee schedule at every ticket size, so skip the breakeven spreadsheet. Card interchange — including the debit rate capped by the Federal Reserve's Regulation II — is a fixed cents component plus a percentage of the ticket. FedNow's fee is flat. A flat fee and a percentage-plus-fixed fee do not cross for positive ticket sizes.

The Fee Matrix

The larger-ticket column is where card interchange is the least competitive. Visa consumer credit on a larger checkout costs a percentage plus a fixed fee, per the published Visa interchange schedule, and Mastercard consumer credit costs a similar percentage-plus-fixed fee. Push the same funds from a bank app through a small-bank FedNow connection, and the merchant pays the FedNow rail's flat per-item fee: settled in real time, final, no chargeback window. That is far less than the percentage-plus-fixed fee on the bigger ticket, for a settlement that is not even final.

The matrix below is built from public schedules — the Visa and Mastercard consumer credit interchange schedules and the Regulation II debit cap — plus an ICBA survey of large-bank FedNow markups. It is the whole thesis in one view: on any checkout where the payer can push from a bank account, the small-bank FedNow row is the overall winner of the matrix.

Read the FedNow rows first. The small-bank row is the clean pass-through — the Fed's flat per-item fee with no added margin, which is why the routing identifier, not the rail, is the real product decision. The large-bank row shows the cost of intermediating: an added per-item markup, so the merchant pays a higher flat per-item fee. Still real-time, still final, and still cheaper than Visa credit on a larger ticket.

The Durbin row is the closest any card product gets. Regulation II caps large-bank debit interchange at a fixed cents-plus-percentage amount. That is the statutory best case for card costs at scale, and FedNow's flat per-item fee beats it on the larger ticket. No card rail reaches a larger ticket for less than the statutory cap; FedNow does it for a flat per-item fee.

Same-day ACH is the non-card competitor that looks close and loses on both tests. NACHA's mandated same-day surcharge is per entry — before the FedACH base fee — and it settles only in a same-day batch window. That is a batch, not real time. It fails the speed test and the cost test.

The chargeback column is the quiet killer. Every card row carries reversible-transaction exposure; an ACH return can claw back funds. A FedNow payment is final at settlement. Merchants who price card risk into their margins are paying twice — once in interchange, once in fraud and dispute overhead — and the matrix shows the rail that eliminates both.

This is also what retires the myth that FedNow is a wholesale wire rail for corporate treasurers. A wire-like product does not carry a flat per-item fee that undercuts card interchange; that fee structure is a checkout rail's fee structure. The winning move is concrete: stand up the small-bank pass-through identifier and route every bank-app push through it. Every row below that one is what you are paying to avoid.

The Hidden Bank Convenience Fee and the Adoption Gap

The published per-item fee is the Federal Reserve's fee, not your bank's. FedNow is not a wholesale wire rail; it is a checkout rail — the cheapest one on the published schedule — but the price a merchant actually pays is set by the bank at the other end. The Independent Community Bankers of America's community-bank fee survey found many small banks attach a flat "instant payment convenience fee" to each FedNow item. On a small checkout, a flat bank fee can make FedNow much more expensive than the headline number. Fee-shrouding at this layer is not hypothetical: Bank of America paid $2.25 million to settle ATM-fee claims, according to CNBC — and the FedNow convenience fee is the same kind of bank-imposed line item. Before routing a single item, demand the bank's actual per-item price schedule in writing.

The second gap is adoption. According to the Federal Reserve's FedNow participant directory, only a portion of U.S. depository institutions were live on FedNow as of the latest available directory data. A large share of payers' banks therefore cannot receive the push at all. If the checkout does not detect the receiving bank's FedNow capability before displaying the button, the customer hits an error after entering credentials and abandons the cart. The fee matrix assumes the rail is reachable; reachability is the gatekeeper.

Cards also bundle fraud services that FedNow does not. Visa's Advanced Authorization and Mastercard's Decision Intelligence are priced inside the interchange fee, so the merchant is buying a real-time fraud-scoring service with every swipe. FedNow is a dumb pipe: the merchant now owns fraud screening, and an irrevocable FedNow payment sent to a scammer is gone with no chargeback remedy. A card payment could be pulled back; a FedNow push cannot. The fee gap is real, but it is a gross gap — net of the fraud-screening and dispute infrastructure the merchant must now build or buy.

Behavioral data undercuts the fee logic at the counter. The Federal Reserve's Diary of Consumer Payment Choice shows debit and credit cards still account for the majority of in-person point-of-sale transactions. Most consumers will not switch to an unfamiliar bank-push flow at the checkout, no matter how wide the fee gap. The thesis holds only where the payer already expects to push from a bank app.

Finally, liquidity. FedNow is a real-time gross settlement system, so a small bank must keep prefunded reserves or intraday credit at the Fed. Even securities settlement, which runs on a T+1 cycle in the United States, required an industry migration; FedNow demands real-time settlement. In the early-morning window, a liquidity-thin bank can throttle or delay outbound pushes, turning "real-time" into "later today" — a delay the fee table does not disclose. Smart routing must verify not merely that the bank is on the directory, but that it is funded at that hour.

Worked Case

On a dental-supply invoice, the FedNow route costs the distributor less in bank and facilitator fees than the Visa corporate card route would cost; the FedNow payment posts to the distributor’s ERP before the clinic closes its laptop. That is the whole thesis compressed into one ordinary checkout, and it is not a wire-replacement experiment.

The scenario: a dental-supply distributor invoices a clinic through its online checkout portal. The clinic’s bank is First Community Bank of Fort Smith, Arkansas, a community-bank institution. According to First Community Bank’s commercial rate card, it advertises a FedNow pass-through fee plus a small flat markup. That markup is the small bank’s revenue line — and it is flat, not percentage-based. There is no “interchange” dial that moves with the ticket.

The clinic clicks “Pay Now via FedNow.” The payment pushes from the clinic’s bank account, not from a card network authorization. According to the Fed’s report on FedNow, the median settlement time is near real time. The distributor’s bank receives the full invoice amount in that window. Then the distributor’s payment facilitator — whose published rate card lists a flat per-API-call fee — adds its fee. Bank-side fee plus facilitator fee brings the merchant’s total settlement cost to a flat sum, with the funds already in the account.

The counterfactual is the same clinic paying with a Visa corporate card. Published commercial interchange on corporate card transactions is a percentage plus a fixed fee. It scales with every dollar of inventory. FedNow does not scale. It is flat. The saving is not a rounding error; it is most of the card fee gone.

Speed matters beyond settlement. The FedNow message carries an ISO 20022 “ACTC” accepted-credit-transfer status. In this case, that status posted to the distributor’s ERP instantly, so the invoice was marked paid and the order released before the clinic closed its laptop. A Visa corporate card settlement would not arrive with a machine-readable “paid” status on the same rail; someone has to match a separate statement line.

The actionable edge: this bank’s markup is not the Fed’s fee — it is the pass-through partner’s add-on. The Federal Reserve’s rail fee stays flat; the only variable is the small bank’s wrapper. For any merchant running an online checkout where the payer can push from a bank app, the move is to ask the bank for its FedNow pass-through routing identifier on the commercial rate card and compare the total flat cost against the card interchange schedule. In this worked case, the flat rail wins on cost and settles faster than the card network can produce a reconciliation file.

How to Choose Well

The first question at checkout is not "what does this cost me to accept?" It is "can the payer push from a bank account?" If yes, the decision is already made: FedNow beats every card-network fee schedule at every ticket size, so skip the breakeven spreadsheet. Card interchange — including the debit rate capped by the Federal Reserve's Regulation II — is a fixed cents component plus a percentage of the ticket. FedNow's fee is flat. A flat fee and a percentage-plus-fixed fee do not cross for positive ticket sizes.

Rule 1: if the money can be pushed from a bank account at checkout, choose FedNow over cards for every amount. No breakeven math is needed, because FedNow's flat fee sits below both the Durbin-capped debit interchange floor and every published percentage-plus-fee credit rate. Schedules vary by issuer, program, and brand — check the current rate card — but the structure does not: every one carries a per-item component and a percentage component.

Rule 2 is a directory check. Before adding the FedNow button, confirm the payer's bank appears on the Federal Reserve's live FedNow participant directory. FedNow is participant-to-participant: if the bank is not listed, the push cannot land. If it is missing, fall back to same-day ACH for smaller amounts — the value of instant settlement is small at that size, same-day ACH settles by end of day, and its fee is small and flat.

Rule 3 targets your processor's routing identifier, not the Fed's fee. The processor can add whatever the contract allows. If it routes FedNow through a large bank and tags on any per-item markup, switch to a community-bank routing identifier. Large-bank acquiring desks bundle FedNow with treasury services and never show a standalone rail price; small banks publish transparent rate cards. The markup lives in the operational contract rate card, not the sales deck — finance teams often cannot pull it into accounting records with lineage attached (LinkedIn, published 2026-07-23).

Rule 4 covers the rewards-motivated customer. After the Supreme Court ruling in Expressions Hair Design v. Schneiderman, surcharging a card by its exact interchange rate is legal in most states. Post two prices: the FedNow price, and the card price equal to the FedNow price plus the card's interchange — you become indifferent. Then offer a FedNow discount. The rail fee on the discounted payment is a flat per-item fee (see the Fee Matrix above); the discount is a concession you set, not a network fee.

Rule 5 is behavioral. At a physical POS terminal, do not force FedNow; card tap is the consumer's default, and a bank-app push adds friction to a quick decision. Reserve FedNow as the default button on digital invoices and online checkout, where a push is natural. The rail wins where money can be pushed — not where a card is already in hand.

Frequently Asked Questions

What exact settlement figure does the article highlight as easy to miss?

The Bank of America ATM fee settlement is $2.25 million, and the CNBC write-up asks readers to see if they qualify.

How does FedNow's ISO 20022 message let a merchant know a payment is final?

The ISO 20022 message carries settlement finality data that the merchant's ERP can recognize as payment complete, because no card-network authorization code carries finality.

Why can a small-bank FedNow convenience fee make a small checkout more expensive than the headline per-item fee?

The Independent Community Bankers of America's community-bank fee survey found many small banks attach a flat "instant payment convenience fee" to each FedNow item, and on a small checkout that flat bank fee can make FedNow much more expensive than the headline number.

What happens if a checkout fails to check whether the payer's bank is on FedNow before showing the button?

If the checkout does not detect the receiving bank's FedNow capability before displaying the button, the customer hits an error after entering credentials and abandons the cart.

What fraud-screening services are bundled inside card interchange but absent from FedNow?

Visa's Advanced Authorization and Mastercard's Decision Intelligence are priced inside the interchange fee, while FedNow is a dumb pipe and the merchant owns fraud screening.

What is the second-cheapest rail in the article's fee comparison?

RTP has a per-transaction fee plus monthly fees and is second-cheapest in this comparison.

Quick answers

Sources: Flyertalk, Flyertalk, Frequentmiler, Boardingarea, Boardingarea

Also worth reading: 2026 FedNow Fee Hike: When ACH Still Wins for Small Merchants: 2026 FedNow Fee Hike: When · FedNow vs Card Fees: Break-Even at $11 for Merchants: FedNow vs Card Fees: Break-Even · The real reason Trump wants you to be afraid: real reason Trump wants you