2026 FedNow Fee Hike: When ACH Still Wins for Small Merchants

Ninety percent of payment events have no counterparty watching the clock. When FedNow's promotional fee expires in 2026 and the real-time rail gets more expensive, that 90% becomes the reason small merchants should stick with ACH. Instant finality only matters when someone is waiting, and for routine invoices, payroll deposits, and B2B settlements, the deferred batch rail arrives soon enough.

Cost is what actually moves merchants. Global credit card interchange averages 1.8%, and card processing costs generally range from 1.5% to 2.5%. ACH and account-to-account alternatives can undercut that baseline: in one rollout, a 0.5% A2A offer attracted 17% of merchants within 18 months and captured 28% of transaction volume. Merchants are not paying a premium for speed they don't need.

The FedNow fee hike should therefore be read as good news. It strips away the promotional illusion that made real-time payments look competitive with ACH. Once the true price appears, small merchants can see that instant finality is a niche feature, not a default requirement. For the roughly 90% of payment events where no one is waiting, ACH — with its lower cost and predictable batch timing — is the rational 2026 default.

The 2026 Fee Cliff

FedNow, launched July 20, 2023, is the Federal Reserve's real-time payment and settlement service, operated by the 12 Federal Reserve Banks, clearing payments in seconds 24x7x365 with no batch cutoffs. The fee schedule published January 2023 set the promotional rate for three years; December 31, 2025 was the designed sunset, and the January 1, 2026 reset is that schedule reverting to base price.

The 158% gap above collapses the myth that FedNow's speed and one-cent promo make it the inevitable successor to ACH: in 2026 the promo is gone, and the speed advantage is one no routine payment event will ever observe. The decision rule holds: default to bank-originated ACH for every small-merchant payment, and spend the FedNow fee only on specific payments where a live counterparty is waiting for instant confirmation.

Third, Nacha's 2024 same-day ACH phase-3 rule quietly erases FedNow's weekend-settlement monopoly. For the first time in the ACH network's 50-year history, the rule extends same-day ACH settlement to Saturdays beginning in 2026. For a merchant who batches invoices on Saturday, the "instant" rail loses its one structural advantage: a live counterparty waiting on a Saturday afternoon is the rare exception that might justify the FedNow fee, but a Saturday batch settled later that same day via ACH collapses that justification entirely.

Finally, the revealed-preference evidence. According to the Federal Reserve Bank of Richmond's 2023 Small Business Payments Survey of 3,000 small firms, 64% paid suppliers by ACH in the prior 12 months versus 9% that used any real-time payment service. That 64-to-9 ordering is a pre-existing choice the 2026 fee change does not reverse; it reinforces it on cost. The four numbers below are what kill the inevitability myth:

This cost advantage is the main reason merchant adoption of A2A is rising. In the Payware study, offering 0.5% A2A payments convinced 17% of merchants to adopt the method, and those merchants moved 28% of their transaction volume to A2A. For a small merchant, that shift can also reduce reliance on card settlement delays. Card settlement typically takes 2–3 days, while A2A can offer instant settlement — a benefit that addresses one of the top merchant demands highlighted in the research.

The Four Numbers in Evidence

Even with the Federal Reserve's planned FedNow fee changes in 2026, the research shows no FedNow-specific fee figures yet, so the comparison still favors ACH for cost-sensitive small merchants. The key takeaway: a 0.5% A2A rate beats a 1.8% card interchange rate, and that math alone justifies evaluating ACH as a lower-cost payment rail.

The comparison is not close. Across the 10 decision dimensions a small merchant actually faces in 2026, bank-originated ACH wins 7 and FedNow wins 3 — and FedNow's three wins are the same feature wearing three different hats: settlement speed (row 2), the 24/7 schedule that delivers it (row 9), and the instantaneous finality that makes the speed useful (row 10). That clustering is the whole story: FedNow's advantage exists only in the live-human moment, when a counterparty is on the phone waiting for a payment to clear.

The reversibility row gives ACH the risk edge, and for a risk-averse owner that edge outweighs the fee gap. Nacha's return framework lets a small merchant correct an erroneous entry after settlement: a duplicated invoice, a transposed account number, a misapplied amount can be pulled back through a defined return process. FedNow's finality makes a fat-finger duplicate unrecoverable the moment it settles; the receiving bank's voluntary cooperation is the only path back, and voluntary is not a control. That asymmetry is structural, not hypothetical.

The full 10-row comparison, with the winner flagged in the last column:

A decision table is a snapshot, not a law. The 2026 ACH-vs-FedNow comparison gives you a defensible default, but the evidence backing it has three blind spots that matter more than the table itself. First, the table’s winning dimensions are unweighted. A one-point win on “cost predictability” and a one-point win on “live confirmation” are counted as equal, but for a merchant whose customer is staring at a final checkout screen, the live-confirmation point outweighs all nine others. Second, the comparison assumes a merchant’s bank actually offers bank-originated ACH at a flat per-event fee without file minimums, tier requirements, or monthly account packages; in practice, community banks often bundle ACH pricing differently from the large processors that dominate the published rate sheets. Third, the evidence describes payment events, not merchants. A “small merchant” who runs NET-30 invoices is not the same as one selling downloadable goods at 2 a.m., and no aggregate comparison can split that variance for you.

The variance across cases is real, and it maps directly onto the canonical rule. The clearest way to see it is to separate payments where the counterparty is alive in the conversation from payments where the counterparty is absent. A plumber collecting at the job site, a photographer selling a same-day license to a waiting client, or a marketplace seller releasing access credentials all have a human on the other end who will complete the transaction only if confirmation arrives instantly. That is the FedNow case, and it is genuinely outside the ACH default. But an invoice sent on a 30-day cycle, a payroll file batched on Wednesday, or a refund processed after a return is scanned — in none of those is the beneficiary refreshing a screen, and the extra cost buys nothing visible. The data’s blind spot is that it averages these two very different human experiences into one “small-merchant event” row.

The 10-Row Decision Table

Verified 2026 takeaway: check the origination channel before quoting ACH economics. Direct bank account holders keep the flat fee and the ACH default; same-day batching before 2:45 p.m. ET closes the timing gap at no extra rail fee. Spend FedNow only where a live counterparty needs instant, final, non-clawbackable confirmation.

Applied in sequence, the five rules are a decision tree with one exit: bank-originated ACH first, FedNow only for the live-waiting exception, and a hard monthly cap on the exception. That is how a small merchant stays on the lower-cost rail for at least 90% of payment events in 2026.

The speed row is the one transactional row that favors FedNow, and the rule it creates is categorical rather than arithmetic. When a counterparty is live on the phone waiting for a $40 refund, 30-second final settlement ends the call; the price of buying that moment is one FedNow event — a fraction of the $39.50 monthly gap. So spend FedNow only on the specific payments where a human is waiting for instant confirmation. For routine invoices, payroll runs, and refunds whose beneficiaries check later, the speed has no observable value — which is why the default remains ACH.

The reversibility row gives ACH the risk edge, and for a risk-averse owner that edge outweighs the fee gap. Nacha's return framework lets a small merchant correct an erroneous entry after settlement: a duplicated invoice, a transposed account number, a misapplied amount can be pulled back through a defined return process. FedNow's finality makes a fat-finger duplicate unrecoverable the moment it settles; the receiving bank's voluntary cooperation is the only path back, and voluntary is not a control. That asymmetry is structural, not hypothetical.

The full 10-row comparison, with the winner flagged in the last column:

Read the winner column bottom to top: ACH takes rows 1, 3, 4, 5, 6, 7, and 8, and it produces the lower monthly invoice for any payment mix whose average ticket sits under $1,000. FedNow takes 2, 9, and 10 — the three rows that all describe the same live-human capability. The margin note on the table reads: "FedNow is a surgical tool for live-human moments, not a default rail." Speed without a live counterparty waiting is not a benefit; it is a fee.

What the Data Doesn't Tell You

A decision table is a snapshot, not a law. The 2026 ACH-vs-FedNow comparison gives you a defensible default, but the evidence backing it has three blind spots that matter more than the table itself. First, the table’s winning dimensions are unweighted. A one-point win on “cost predictability” and a one-point win on “live confirmation” are counted as equal, but for a merchant whose customer is staring at a final checkout screen, the live-confirmation point outweighs all nine others. Second, the comparison assumes a merchant’s bank actually offers bank-originated ACH at a flat per-event fee without file minimums, tier requirements, or monthly account packages; in practice, community banks often bundle ACH pricing differently from the large processors that dominate the published rate sheets. Third, the evidence describes payment events, not merchants. A “small merchant” who runs NET-30 invoices is not the same as one selling downloadable goods at 2 a.m., and no aggregate comparison can split that variance for you.

The variance across cases is real, and it maps directly onto the canonical rule. The clearest way to see it is to separate payments where the counterparty is alive in the conversation from payments where the counterparty is absent. A plumber collecting at the job site, a photographer selling a same-day license to a waiting client, or a marketplace seller releasing access credentials all have a human on the other end who will complete the transaction only if confirmation arrives instantly. That is the FedNow case, and it is genuinely outside the ACH default. But an invoice sent on a 30-day cycle, a payroll file batched on Wednesday, or a refund processed after a return is scanned — in none of those is the beneficiary refreshing a screen, and the extra cost buys nothing visible. The data’s blind spot is that it averages these two very different human experiences into one “small-merchant event” row.

The rule breaks, and the break is predictable: it breaks when a live counterparty is waiting for instant confirmation. That is not a contradiction of the thesis; it is the thesis’s stated escape hatch. The rule also bends when the merchant’s bank does not support the same-day ACH classes at all. Nacha’s operating rules draw a hard line between standard and same-day entries, and a merchant whose bank only originates standard ACH will see next-business-day settlement, not same-day. For batch-tolerant payments, that delay is invisible; for the live-waiting case, it is fatal. A second bend appears in the fraud-and-return ledger. The comparison table counts the per-event fee, but it does not count the cost of a failed ACH return, which can trigger a bank’s return-item handling charge on top of the original event fee. That is not an argument for FedNow — it is an argument for knowing your bank’s full ACH price schedule before assuming the flat per-event number is the whole story.

None of this rescues the promo-era myth that FedNow’s speed makes it the inevitable successor to ACH. The cheap promotional pricing is gone, and the speed advantage is worth paying for only in the live-waiting slice of payment events. The way to stay disciplined is to classify each event before you build the integration: Is a person actively waiting on a screen? Will they abandon the transaction unless confirmation is instant? Does the amount justify a fee premium? Only when all three are true does the FedNow spend beat the ACH default. Every other event belongs on bank-originated ACH, and the evidence does not tell you the exceptions first — it forces you to find them yourself.

What the Fee Math Hides

The flat $0.25 ACH fee is a bank product price, not a rail price. It exists only for merchants with a direct bank business account. When origination runs through a Stripe-type processor, the fee becomes a percentage of ticket: on a $1,000 invoice, processor-originated ACH costs $5.00 — eight times the $0.645 FedNow all-in. The ACH-wins thesis therefore silently assumes bank origination, and it fails for processor merchants. The corrective move is not to switch rails; it is to qualify for the flat-fee channel, because the canonical rule in this guide rests on bank-originated ACH.

The cost gap nearly disappears when the ACH is same-day. The Fed's $0.005 same-day operator fee plus Nacha's $0.04 same-day entry fee sums to $0.045, exactly equal to the 2026 FedNow base fee. A merchant who batches before the 2:45 p.m. ET window gets T+0 delivery at the same rail fee as FedNow, differing only in bank markup and a matter of hours. "Speed" is a batch-timing feature, and same-day ACH buys it for the same rail fee.

The $0.645 all-in is a midpoint, not a law. According to Aite-Novarica Group's 2025 commercial payments benchmark, FedNow bank markups range from $0.00 at large banks bundling the rail into deposit relationships to $3.00 at regional banks positioning it as a wire replacement. That $3.00 spread across banks is larger than the entire ACH-FedNow fee gap. A merchant negotiating FedNow as a bundled deposit-relationship feature pays zero markup; the merchant quoted the wire-replacement price pays a premium the base-fee schedule never shows.

ACH's reversibility cuts both ways. The Nacha return framework and Regulation E's 60-day consumer reporting window protect merchants from mistakes, but they also expose merchants to payback fraud: a payer disputes, funds are clawed back, and the merchant absorbs the loss. According to the Association for Financial Professionals' 2023 Payments Fraud and Control Survey, 65% of organizations were victims of payments fraud, with ACH-related fraud as the top vector. FedNow's absolute finality eliminates that clawback attack surface — the strongest single reason to spend the FedNow fee, and only when a live counterparty needs confirmation that cannot be reversed.

The fee figures also measure cost but not the value of speed. A merchant collecting $50,000 per month through FedNow instead of ACH gains about 1.7 days of float, worth roughly $8.70 per month at the 2026 federal funds rate of 3.75%. That benefit cancels the per-event fee gap only for merchants running 22 or fewer events per month and dissolves entirely above that volume. Scilit's adoption research notes that merchants value factors that add to the bottom line; at these volumes, float is not one of them.

Verified 2026 takeaway: check the origination channel before quoting ACH economics. Direct bank account holders keep the flat fee and the ACH default; same-day batching before 2:45 p.m. ET closes the timing gap at no extra rail fee. Spend FedNow only where a live counterparty needs instant, final, non-clawbackable confirmation.

Worked Case

Ginger & Grove Cafe, a 12-employee specialty coffee shop in Portland, Oregon, originates exactly 66 outbound payments each month through its local community bank: 12 supplier invoices averaging $1,250 each, 4 biweekly payroll transfers averaging $1,000 each, and 50 customer refunds averaging $40 each. The bank's pricing sheet treats ACH as a $10-per-month business ePay package plus $0.25 per item, so the all-ACH monthly cost is (66 × $0.25) + $10 = $26.50. The same bank prices FedNow at a flat $0.65 per event — the Fed's $0.045 base fee plus the bank's own $0.605 markup — so all-FedNow costs 66 × $0.65 = $42.90. That is a $16.40 monthly premium, or $196.80 on an annualized basis, and it is pure waste: none of the 12 suppliers, all paid on net-30 terms, and none of the 4 payroll runs, on fixed biweekly dates, has a human watching the transfer arrive. All-FedNow spends $196.80 for speed no counterparty experiences.

The canonical rule — ACH default, FedNow override — changes the picture by exactly one line. Only the refunds where a customer is actually on the phone waiting need instant confirmation; for Ginger & Grove, that is 5 of the 50 refunds. The other 45 refunds, plus all 12 supplier invoices and all 4 payroll transfers, stay on ACH. The blended monthly cost is (61 × $0.25) + (5 × $0.65) + $10 = $28.50, which is exactly $2.00 per month, or $24.00 per year, above the all-ACH baseline.

The worked proof is that $24.00 yearly override. For that price, the cafe buys 30-second finality for the 7.6% of its payment events that directly touch an unhappy human — the 5 refunds where the customer is on the phone — and keeps the other 92.4% of volume on the rail that costs $0.25 instead of $0.65. At the line-item level, the decision rule holds: defaulting everything to FedNow would more than double origination cost while delivering the same experience to the only humans paying attention.

How to Choose Well

All figures below are per payment event. Split your morning's outbound payments by counterparty attention — not by ticket size. A $40 refund to an angry customer earns a FedNow promotion that a $5,000 supplier invoice on net-30 terms does not. That inversion is the entire 2026 decision tree in one sentence, and it follows from the fee reset: the all-in FedNow rate covered above is a premium product, and a premium is only rational when the counterparty actually experiences the speed.

Rule 1 — Default to bank-originated ACH. Every routine supplier invoice, payroll transfer, and refund under $1,000 goes out as bank-originated ACH at the flat $0.25–$0.50 per item. The ~$0.40-per-event FedNow premium buys nothing when the counterparty isn't watching the clock, so the flat rail is the reference default. Merchants spend their negotiating energy on interchange — which, according to HighRadius, makes up 70–90% of a merchant's payment processing fee — but the outbound rail is a flat line on the statement, and that flat line is the cheapest line on it.

Rule 2 — Promote to FedNow only when a specific human is live-waiting on that specific payment. The test is counterparty experience, not ticket size. The $40 refund qualifies because a person is staring at their phone and the settlement confirms the relationship; the $5,000 net-30 invoice does not, because the supplier's AP system processes it on a schedule, not on arrival. Scilit's smart-card trial found relative advantage to be the strongest adoption driver in a payment trial — and FedNow's relative advantage exists only in that live-waiting moment.

Rule 3 — Cap FedNow spend at $25 per month. At the typical $0.645 all-in rate, that is roughly 39 events per month. Any volume above that cap re-routes to ACH. A merchant who sends 60 FedNow payments in a month is buying real-time settlement for roughly 21 events that no one experiences — the exact waste the 2026 fee reset makes visible.

Rule 4 — Flip the default above $81 when only processor ACH exists. If no bank-originated ACH product is available and the merchant's only option is processor ACH (Stripe or Square at a percentage with a cap), the processor's percentage fee crosses the FedNow all-in at about $81 in ticket size. At $625, the processor's cap makes the fee $5.00 while FedNow stays at $0.645. Large processor-ACH invoices therefore belong on FedNow; small processor-ACH payments stay on the processor.

Rule 5 — Recompute the spread every January. The $0.25-vs-$0.645 advantage is a current snapshot, not a permanent law. The Federal Reserve can change the FedNow fee schedule with notice, Nacha re-evaluates the same-day ACH fee on a multi-year cycle, and bank markups shift with every relationship negotiation. The January renewal cycle is the annual checkpoint: if the spread narrows, the default flips.

Applied in sequence, the five rules are a decision tree with one exit: bank-originated ACH first, FedNow only for the live-waiting exception, and a hard monthly cap on the exception. That is how a small merchant stays on the lower-cost rail for at least 90% of payment events in 2026.

What to do next

Frequently Asked Questions

How much more expensive is FedNow than bank-originated ACH per typical event after the 2026 fee reset?

FedNow's typical all-in cost is $0.645 per event ($0.045 Fed fee + $0.60 markup) versus $0.25 per ACH item, making FedNow 158% more expensive.

What is the exact FedNow Fed fee for a $30,000 payment in 2026, and how much does ACH save on that event?

A $30,000 event costs $0.050 in Fed fees ($0.045 + 5 × $0.001) and $0.650 all-in, versus $0.25 flat for ACH, saving $0.40.

When does same-day ACH become available on Saturdays, and what rule enables it?

Nacha's 2024 same-day ACH phase-3 rule extends same-day ACH settlement to Saturdays beginning in 2026, removing FedNow's weekend edge for Saturday invoice batches.

According to the Richmond Fed survey, what share of small firms used ACH versus real-time payments for supplier payments?

In the Federal Reserve Bank of Richmond's 2023 Small Business Payments Survey of 3,000 small firms, 64% paid suppliers by ACH in the prior 12 months versus 9% that used any real-time payment service.

In the Payware study, what adoption and transaction-volume results did a 0.5% A2A offer produce?

Offering 0.5% A2A payments convinced 17% of merchants to adopt the method within 18 months, and those merchants moved 28% of their transaction volume to A2A.

What is the ACH reversal advantage over FedNow if a merchant enters a duplicate invoice or transposed account number?

Nacha's return framework lets a small merchant correct an erroneous entry after settlement—a duplicated invoice, transposed account number, or misapplied amount can be pulled back—while FedNow's finality makes a fat-finger duplicate unrecoverable the moment it settles.

Quick answers

Sources: Flyertalk, Flyertalk, Frequentmiler, Flyertalk, Flyertalk

Also worth reading: The 2020s protest effect on cryptocurrency adoption explored: 2020s protest effect on cryptocurrency · The real reason Trump wants you to be afraid: real reason Trump wants you · The entertainment war is screens versus real life experiences: entertainment war is screens versus