FedNow Averages $99K Per Transaction. Is AP Ready for Instant Rails?

Everyone building a multi-rail AP strategy in 2024 made the same assumption. RTP would carry the high-value B2B traffic because The Clearing House had a five-year head start with the big banks. FedNow would sit downstream as a consumer and small-business rail, useful for tail-end supplier coverage and gig-economy payouts.

The Q1 2026 Federal Reserve data says the opposite is happening. In Q1 2026, RTP processed 128 million transactions for $480 billion per quarter, averaging $3,750 per transaction, while FedNow had an average value of $99,414 per transaction (Source: Federal Reserve Bank of Richmond, https://www.richmondfed.org/publications/research/economic_brief/2026/eb_26-15).

That is not a rounding difference. FedNow's average ticket is roughly 26 times larger than RTP's. Whatever theory your AP team built its instant-rail strategy on, that ratio should force a rewrite.

The Frame: Rails Are Not Interchangeable Even When They Look Identical

The temptation with instant payments is to treat RTP and FedNow as substitutes. Same speed. Same finality. Same ISO 20022 message format. Different operator. Route to whichever one your bank supports and move on.

That model breaks the moment you look at how the two rails are actually being used in production.

RTP is behaving like a high-frequency, low-value settlement layer. Payroll top-ups, insurance disbursements, marketplace payouts, gig-worker settlement. The volume tells you it is a rail optimized for millions of small movements.

FedNow is behaving like a wholesale corporate settlement rail. A hundred-thousand-dollar average implies invoice payments, treasury sweeps, real estate closings, and inter-company transfers. The volume is smaller, but each transaction carries the weight of a wire.

Same technology. Completely different risk surface. And the AP teams that route to either rail based on availability rather than intent are underpricing that difference.

Why the User Mix Diverged

Look at the participant data. FedNow has attracted 1,725 banks and credit unions, representing 19.7 percent of U.S. financial institutions as of Q1 2026 (Source: Federal Reserve Bank of Richmond, https://www.richmondfed.org/publications/research/economic_brief/2026/eb_26-15). That skews heavily toward community banks and credit unions who joined FedNow first because the Fed was the operator they already trusted.

Community banks serve middle-market businesses. Middle-market businesses cut six-figure checks. When those businesses got instant-rail access for the first time, they didn't use it to send $50 to a contractor. They used it to settle the invoices that were previously going out by wire or ACH.

Meanwhile, 70 percent of RTP participants also maintain FedNow membership, and 49 percent of FedNow members have joined RTP (Source: Federal Reserve Bank of Richmond, https://www.richmondfed.org/publications/research/economic_brief/2026/eb_26-15). The overlap is meaningful, which means for a growing set of banks, the routing choice is a policy decision, not a technical one. And their corporate customers are voting with volume for FedNow on high-value items.

Whatever combination of endpoint reach, pricing, and treasury workflow drove that split, the outcome is the one that matters for AP. If you are sending a six-figure supplier payment on an instant rail, the odds are better than even it lands on FedNow.

Where the Playbook Breaks

Most AP fraud-control frameworks were written when instant meant small. That assumption is now backwards on FedNow, and it breaks four things in a typical control environment.

Approval thresholds keyed to rail, not amount. A common design pattern is to allow instant payments up to some ceiling with a single approver, and require dual approval above that. If your average FedNow transaction is closing in on six figures, most of your instant-rail volume is now sitting in the exception queue. The workflow was designed for RTP's transaction profile. It does not fit FedNow's.

Sanctions and beneficiary screening tuned for speed over depth. ACH gives you the illusion of screening time. Wires force you to slow down. Instant rails have neither cushion. If you are pushing near-six-figure averages through a rail that clears in seconds and is irrevocable on receipt, the screening logic upstream has to be pre-computed, not real-time. Vendor validation has to happen at master-file entry, not at the moment of payment.

Recall assumptions inherited from ACH. ACH lets you request a return. Wires give you a small window to recall before the beneficiary bank credits. FedNow and RTP do not. Money moves, finality settles, done. Any AP process that has "we can claw it back if something looks wrong" written into its response plan needs to be rewritten for instant rails, especially at the amounts FedNow is now carrying.

Rail selection logic based on cost, not risk. Most treasury teams route instant payments to whichever rail is cheaper per transaction. When those payments were small, the cost delta mattered and the risk delta didn't. When they are averaging close to six figures, the risk delta swamps the cost delta. Routing logic should reflect that.

What Changes When Instant Becomes Enterprise-Grade

Here is the operational reframe. Stop thinking of FedNow as "RTP's sibling." Start thinking of FedNow as a wire replacement with a different failure mode.

Wires clear the same day but not instantly. FedNow clears in seconds. That collapses the window in which any downstream control can operate. And it does it at a transaction size where a single mistake is material to a monthly close.

That reframing changes four things in how a mid-market AP shop should design its instant-rail program.

Vendor validation becomes a gate, not a step. The bank account on the vendor master must be validated, ownership-confirmed, and change-controlled before it becomes eligible for instant-rail routing. Not at payment time. At master-file entry. This is why we built supplier management as a validation layer rather than a directory. The account information behind an instant payment cannot be a field a coordinator edited last Tuesday.

Rail selection becomes risk-tiered. For payments under a defined threshold, RTP is fine. It matches the risk profile of the volume it is already carrying. Above that threshold, either route to FedNow with heightened controls or route to a virtual card, which is revocable, controlled at the credential level, and comes with monetization on top. Virtual card payments are the counter-argument to instant rails for high-value B2B, and the math often wins.

Fraud protection has to be underwritten, not implied. 79% of organizations experienced attempted or actual payments fraud in 2024 (Source: AFP 2025 Payments Fraud and Control Survey, https://www.afponline.org/publications-data-tools/reports/survey-research-economic-data/Details/payments-fraud). At near-six-figure averages on an irrevocable rail, "we have controls" is not a policy. It is a hope. Finexio Shield carries a $2M fraud guarantee. That kind of underwriting is what an enterprise-grade instant-rail program should be paired with.

Approval thresholds need to be reset by amount, not by rail. If your instant-rail policy still keys single versus dual approval to a threshold that made sense for small instant payments, and your average instant payment is now approaching six figures, the policy is doing nothing. Reset the thresholds against the new distribution.

The Counterintuitive Move: Send Less Volume to Instant Rails, Not More

Here is where we take a position most rail-strategy pieces won't.

The right answer for most mid-market and enterprise AP shops is not to accelerate their move to instant rails. It is to be more selective about what goes on them.

Instant rails solve for one thing: speed of settlement. That matters for a specific subset of AP payments. Emergency vendor payments. Trade-discount capture where the discount window is closing. Payroll corrections. Real estate and closing-adjacent flows. Situations where the supplier's cash conversion cycle depends on receiving funds today.

For the rest of the AP file, which is most of the AP file, the right answer is a virtual card program that pays the supplier on their existing terms, captures rebate, and pushes fraud risk onto the card network rather than into your operating account. That is the case for AP payments as a service and it gets stronger as instant rails move upmarket, not weaker.

Instant rails should carry the payments that need to be instant. Card should carry the payments where monetization and control matter more than speed. ACH should carry the long tail. The three-way split is a portfolio decision, and FedNow's shift to near-six-figure averages is a signal that portfolio should be tighter, not looser.

The Finexio Position

We orchestrate payments across rails. The three-party model is Finexio as the orchestrator, J.P. Morgan Chase as the issuing bank, Mastercard and Visa as the card networks. We route across virtual card, ACH, check, and instant rails including FedNow and RTP. Ten-plus years in market. $75M+ in investment.

The value we bring on the instant-rail question is not "we connect to FedNow." Everyone will eventually connect to FedNow. The value is deciding which of your supplier payments should go there, which should not, and applying the controls that make either choice defensible when a high-value transaction lands on the wrong account.

Finexio Shield sits underneath all of it with a $2M fraud guarantee, because on an irrevocable rail at wire-sized amounts, the underwriting has to be there or the strategy is theater.

FAQ

Should our AP team enable FedNow now, or wait?

Enable the rail. Restrict the use case. The mistake is either turning it off entirely, which leaves you unable to move fast when you need to, or turning it on with the same thresholds you used for RTP, which mis-designs your controls for the new volume profile. Enable it, restrict initial usage to a defined set of high-value scenarios, and expand from there.

Does the FedNow data mean RTP is losing?

No. RTP is doing exactly what an instant rail should do at scale, moving millions of small payments efficiently. FedNow is doing something different, moving fewer, larger payments through community-bank endpoints. Both are winning at different jobs.

How does virtual card compete with instant rails at high-value amounts?

On monetization, control, and revocability. A virtual card at that amount generates rebate, is credential-controlled per transaction, and can be voided if something is wrong. An instant rail generates none of that and cannot be reversed. For supplier payments where speed is not the actual constraint, card wins the comparison. See monetize AP payments for the math.

Book a Consultation

If your instant-rail strategy was designed against earlier assumptions and the 2026 data has changed the picture, this is the right time to revisit it before the next audit cycle catches up. Finexio can walk your team through rail selection, vendor validation, and the portfolio split between card, ACH, and instant. Book a consultation with our team.

Sources

- Federal Reserve Bank of Richmond: https://www.richmondfed.org/publications/research/economic_brief/2026/eb_26-15 - AFP 2025 Payments Fraud and Control Survey: https://www.afponline.org/publications-data-tools/reports/survey-research-economic-data/Details/payments-fraud