Two rails, one customer promise
FedNow and the RTP network are back-end bank-account payment rails, not consumer apps. FedNow launched in July 2023; The Clearing House launched RTP in November 2017. Both operate 24/7/365, use ISO 20022 messages, support credit transfers and request-for-payment messages, and allow transfers up to $10 million. The institution or fintech controls the customer interface, eligibility, lower product limits and use cases, so joining a rail does not by itself create a complete instant-payment product. [1][2][3]
The central design promise is immediate interbank settlement and rapid availability of funds. That is different from an “instant-feel” wallet, card push or account display that may make value available before external settlement finishes. The distinction matters because the provider advancing funds may still carry , fraud or counterparty exposure while settlement catches up. [2]
Speed has different value for different payments
An emergency disbursement, a supplier payment and a routine bill do not create the same urgency. A recipient short of usable cash may value immediate availability highly, while a predictable payment scheduled in advance may work well on another rail. Faster settlement is a capability whose value depends on the problem it solves.
A financial institution can therefore begin with a defined customer need rather than a universal migration target. The comparison should include certainty of receipt, information accompanying the payment and the cost of resolving mistakes. A service that moves money quickly but leaves the recipient unable to identify its purpose can still create substantial reconciliation work.
Final settlement moves the control point forward
For RTP, The Clearing House says a sending institution cannot revoke or recall a payment after submission; settlement is final and irrevocable, although the network includes a message requesting return of funds. For FedNow, Regulation J commentary states that a Federal Reserve credit is final and irrevocable when made. A final rail transfer is therefore not equivalent to a guaranteed recovery from the beneficiary after a mistake or scam. [3][4]
Finality does not erase every legal right. Regulation J expressly preserves the Electronic Fund Transfer Act where it is inconsistent with the FedNow rule, and the commentary addresses unauthorized and erroneous orders. Consumer protections, bank-customer agreements, mistake and restitution law, and network procedures may still matter. The practical control lesson is narrower: authentication, account validation, sanctions screening, velocity checks and confirmation of the payee and purpose should happen before the payment is released. [4]
Request for payment is a message, not a debit mandate
A request for payment can carry invoice or bill information to a payer, but it does not itself move money. Federal Reserve commentary says certain FedNow requests for payment are not treated as payment orders because they do not instruct a Reserve Bank to pay. The payer or its institution must separately authorize the credit transfer. [4]
That separation is useful but not self-protecting. A fraudulent invoice can still persuade an authorized user to approve a valid payment order. Controls should bind the request to the expected biller, show the destination clearly, detect first-time or changed payees, and give the customer a meaningful confirmation step. Treat “authorized” and “not manipulated” as two separate questions.
Working-capital value should be calculated over the actual delay
Suppose a hypothetical business receives $100,000 three days earlier and can avoid borrowing at an assumed 10% annual rate. Using a 365-day simple-interest convention, the financing value is about $82. That is a timing benefit, not $100,000 of new revenue, and it may be smaller than other benefits or costs of the transaction.
For a household, the relevant value may instead be avoiding a missed obligation or obtaining essential goods promptly. Those outcomes cannot be inferred from the business interest calculation. Assess the recipient’s actual situation and any fees rather than treating every accelerated dollar as having the same economic benefit.
Receive-only is not send readiness
Many institutions start by receiving. Receiving still requires posting funds, screening exceptions, keeping the core ledger available and supporting customers outside normal hours. Sending adds a harder decision: the bank must authenticate the customer, evaluate fraud and account risk, confirm available funds, apply product limits and release an instruction that settles rapidly. A directory listing or receive capability should not be described as full send availability.
A March 2025 network case study says ABNB Federal Credit Union joined RTP in May 2024 and worked with Alacriti to progress from receive to send. The case study describes a pre-launch database scrub, biometrics, multifactor authentication and fraud scoring. It also reports more than $1 million received across 3,000 transactions in the first 90 days. These are company and network claims about one implementation, not independent evidence of industry-wide loss rates or profitability. [8]
Liquidity differs even when the customer experience looks similar
The rails use different settlement structures. A Richmond Fed analysis describes RTP as using a prefunded joint account at the Federal Reserve, requiring participants to preposition funds. FedNow settles through a participant’s Federal Reserve master account or a correspondent’s designated settlement account and provides -management transfers. The Clearing House also lists funding agents that can manage RTP settlement accounts and round-the-clock liquidity for smaller institutions. [2][4][7]
Neither design makes liquidity free. Treasury teams need weekend and holiday thresholds, alerts, replenishment authority and a contingency if a correspondent, funding agent or connection fails. A direct participant should also measure the opportunity cost of prefunding; an indirect participant should examine concentration, service-level and contractual exposure to its provider.
Scroll horizontally to see all columns.
| Operating question | FedNow | RTP |
|---|---|---|
| Where settlement occurs | Participant master account or designated correspondent account at a Federal Reserve Bank | Prefunded joint account structure described by the Richmond Fed |
| Smaller-FI access path | Certified service providers, settlement agents and liquidity providers | Third-party service providers and funding agents |
| What must work overnight | Account capacity, connection, posting and exception handling | Prefunding or funding-agent coverage, connection, posting and exception handling |
| Customer protection focus | Pre-send authentication and product controls plus applicable legal protections | Pre-send authentication and product controls plus return-request and network procedures |
A coherent product can use several rails
A provider can offer different payment methods for different timing, reach and recovery needs. The customer should understand when money will arrive and what the choice costs without needing to learn the entire settlement architecture. Internally, the institution must still know which rules and operating dependencies apply.
The 2026 FedNow fee schedule is one input to that comparison, not the full cost of a retail or treasury service. Integration, , support and fraud handling can dominate the rail charge. Measure completed customer use cases and total contribution alongside transaction growth, preserving the distinct FedNow and RTP operating characteristics below. [5]
Adoption is broadening, but the two volume series are not interchangeable
Official FedNow statistics show 4,997,811 settled customer credit transfers totaling $274.664 billion in the second quarter of 2026. Volume rose 83.2% from the first quarter while value rose 1.3%, reducing average value to $54,957 from $99,414. The Clearing House reports 142 million RTP transactions totaling $576 billion in the same quarter and more than 1,357 participants as of August 2026. RTP therefore carried far more transactions in the quarter, while FedNow’s much larger average payment signals a different mix. [1][6]
Do not convert those figures into market share without a common denominator. The rail operators publish different participant and usage measures, participants can join both rails, and an institution may be receive-only or enable only selected customers. The Richmond Fed counted 1,725 FedNow participants and 1,193 RTP participants in its first-quarter 2026 comparison, with 837 on both. It also found that the combined U.S. instant-payment volume still trailed several international systems on a per-capita basis. [2]
Separately, The Clearing House’s RTP overview, checked October 4, 2026, reports 150 million transactions totaling $621 billion for Q3 2026 and more than 1,378 participants as of September 2026. The page does not identify when these figures were posted, and its report button still reads “Q2 2026 RTP Report.” These figures are attributed to the current overview rather than a separately verified Q3 report. The FedNow statistics page checked on the same date still ends at Q2 2026, so the table retains the matched Q2 comparison. [1][10]
Scroll horizontally to see all columns.
| Q2 2026 operator statistic | FedNow | RTP |
|---|---|---|
| Transactions | 4,997,811 settled customer credit transfers | 142 million transactions |
| Value | $274.664 billion | $576 billion |
| Published average value | $54,957 | About $4,056 calculated from rounded operator totals |
| Important limit | Official page excludes non-credit-transfer messages | Operator totals and participant claims are network-reported |
Economics extend beyond the rail fee
The 2026 FedNow fee schedule lists $0.045 for an originated customer credit transfer, $0.01 for a request-for-payment message and $1 for a -management transfer. It waives the general $25 monthly RTN fee and credits the first 2,500 customer transfers per month during 2026. Those are rail charges, not the institution’s total product cost. [5]
A credible business case also includes core integration, fraud tooling, customer support, 24/7 operations, liquidity, vendor fees, reconciliation, compliance and losses. Revenue can come from treasury products, faster disbursements, account retention or explicit fees, but a high transaction count does not establish profit. Measure contribution after fraud, returns, manual work and liquidity—not just the network fee.
Looking ahead, Federal Reserve Financial Services announced on September 15, 2026 that a new FedNow discount program will begin January 1, 2027. It describes statement credits for qualifying receive and send enablement, with receiver discounts linked to specified value-added services and sender discounts subject to baseline activity thresholds. FRFS also says the origination discount for the first 2,500 items per month will end in conjunction with the rollout. These announced incentives concern 2027, distinct from the current 2026 fee schedule. The announcement does not establish a new regulatory requirement or a complete 2027 fee schedule. [11]
A practical approval checklist
Before enabling send, require evidence for five decisions: who may send; how payees and devices are authenticated; which transactions receive manual or automated friction; how 24/7 is replenished; and who owns recovery, complaints and customer communication after an error. Test the whole chain with the core, fraud engine, provider, settlement account and posting process—not only the network connection.
Useful measures include send and receive volume by use case; approval and decline rates; fraud attempts and losses per 10,000 transfers; customer-reported scam rate; rate; time to post; reconciliation breaks; return-request recovery; liquidity alerts; and after-hours support demand. Separate account takeover, authorized-push fraud, mistaken payee and operational duplication because they require different controls.
The assessment would change with verified loss and recovery data by rail and use case, broader evidence on send-enabled institutions, sustained small-value consumer usage, or a material change to network rules and legal treatment. Until then, the strongest conclusion is operational: instant settlement is mature enough to deploy, but readiness is defined by pre-send control, uninterrupted posting and liquidity, and an honest recovery process—not by a logo or participant count.
Sources
- Federal Reserve Financial Services — FedNow Service volume and value statistics; Q2 2026, page updated July 6, 2026SourceBack to text: ↑1↑2↑3
- Federal Reserve Bank of Richmond — FedNow and the Development of U.S. Fast Payments; August 2026SourceBack to text: ↑1↑2↑3↑4
- The Clearing House — RTP network institutional overview and operating characteristics; reviewed September 30, 2026SourceBack to text: ↑1↑2
- Federal Reserve Board — Commentary on Regulation J, including FedNow messages, settlement accounts and finality; reviewed September 30, 2026Official sourceBack to text: ↑1↑2↑3↑4
- Federal Reserve Financial Services — FedNow Service 2026 Fee Schedule; effective January 1, 2026SourceBack to text: ↑1↑2
- The Clearing House — RTP network Q2 2026 volume, value and August participant statistics; reviewed September 30, 2026SourceBack to text: ↑
- The Clearing House — RTP funding-agent function; reviewed September 30, 2026SourceBack to text: ↑
- The Clearing House — ABNB Federal Credit Union instant-payments case study; March 12, 2025SourceBack to text: ↑
- Federal Reserve Financial Services — current FedNow participants and service providers; lists updated September 28, 2026Source
- The Clearing House — RTP overview, Q3 2026 totals and September participants; checked October 4, 2026; posting date not shownSourceBack to text: ↑
- Federal Reserve Financial Services — new FedNow discount program beginning January 1, 2027; announcement September 15, 2026SourceBack to text: ↑