The Stablecoin Banker, April 28

April 28, 2026

Four stablecoin developments bankers should know about

The Stablecoin Banker is a periodic newsletter keeping bankers on top of the stablecoin industry. I highlight top stories that are relevant to banks, with my insights and commentary to draw out the most important conclusions.

If you find this content useful, please feel free to forward it to your friends and colleagues. They can also subscribe directly. If you’re interested in stablecoin services for your bank, feel free to reach out to us at Omnia.

"Let's say you just need a little set of data, like weather data. You can just call an API and that API can be paid one cent instantly. The fiat system is just not geared for that."
— Laurent Lambert, Head of Blockchain, IBM

In This Issue

• Congress takes aim at the bank monopoly on Fed payment rails
• DoorDash and Erebor demonstrate end-to-end stablecoin payroll
• Card networks and crypto protocols race to power AI agent payments
• HSBC plants tokenized deposits in the U.S. — and reveals their limits

Plus, tidbits you may have missed in our Coupon Clippings section.

Congress Tries to Pull Payments Out of the Banking Stack

On April 21, Reps. Young Kim (R-CA) and Sam Liccardo (D-CA) introduced the Payments Access and Consumer Efficiency (PACE) Act, which would create a OCC-regulated nonbank payment service provider license. Under the bill, qualified registrants with 40+ MTLs or a state banking or credit union charter, could apply for a Federal Reserve “payments reserve account” with access to FedWire, FedNow, and FedACH. The proposed bill imposes government review deadlines and mandates 1:1 sequestered reserves for non-settled payments activity. The legislation has been endorsed by the Financial Technology Association, the Blockchain Association, the Digital Chamber, and the Crypto Council for Innovation. Sponsors framed it as a way to reduce bank fees and settlement delays for everyday consumers; industry endorsers framed it as long-overdue access to federal rails for digital asset payment companies. [Sources: Rep. Kim, PYMNTS]

My Take:

PACE is the next step in a nine-month long surge of regulatory and legislative progress on innovation in and outside of banking. GENIUS passed and rules are in flight. De novo digital asset bank charters and non-depository trust charters are being approved in record time. The Fed published an RFI on a limited “payments account” for non-traditional financial institutions. PACE contributes by increasing legislative clarity on what it takes to get access to Federal Reserve payment services - heretofore the province of depository banks, by and large.

However, the bill is best understood as an intention-setting first draft and has some major shortcomings. Notably, the text defines a “payments reserve account” even though the Fed has yet to create such a product. Since the legislative text doesn’t require the Fed to create the product, it’s unclear exactly what the OCC would be granting access to. Furthermore, it conflicts with the Fed’s payments account RFI by including FedACH in this non-existent product. ACH’s 60-day reversibility window creates significant counterparty risk that is not addressed by the bill because it fails to require reserves against this long-tail risk.

Despite its flaws, the intent is clear: build a credible, federal path for non-banks to operate end-to-end on Fed payment rails without a sponsor. GENIUS creates the ability to hold dollars outside of the banking system, and PACE operationalizes it with payments services outside of the banking system.

Banks can react to the increasing momentum towards a more open, heterogenous, and competitive environment. One approach is to focus on customers where a slick payments experience is not a requirement: yield-based savings products, heavy check/ACH users, or deposits tied to commercial loans. This field is narrowing and banks choosing this will need to focus on relentless cost optimization. Banks that have not built a culture that prioritizes and resources product creation as a core capability will be forced down this path.

The second approach is to invest in payments, and build a retail deposit or commercial treasury management product that competes. This is a significant undertaking even for banks with a product culture because their systems architecture drives up costs and slows development. The banks choosing this route will need to allocate material budget and begin thinking about how they can implement it in ways that streamline their overall architecture and reduce their fragmented vendor dependencies.

A third would be to focus on what unique capabilities banks still retain like FDIC insurance, liquidity capabilities and card issuance, and partner with fintechs specifically on those services. This is an attractive path for most banks that have achieved a certain scale and sophistication because it plays to the strengths they already have without requiring a large investment in new product. However, because of its relative simplicity, it will soon become a red ocean with many providers of commoditized bank-services-for-fintechs products.

As the march of regulatory progress continues, banks will face increasing pressure to respond. The second path offers the most opportunity for institutions with growth ambitions over the next decade.

DoorDash Pays Out in Stablecoins. Erebor Bank Takes Them In.

On April 21, DoorDash announced it will route payouts through Tempo, a Stripe- and Paradigm-incubated payments blockchain, alongside production rollouts from Stripe, Coastal Bank, and Latin American payments platform ARQ. DoorDash co-founder Andy Fang said stablecoins offer a way to standardize money flows across the company’s three-sided marketplace in 40+ countries, starting with payout flows. One day later, Infinite launched Infinite Accounts — dedicated bank accounts with unique routing numbers, powered by Erebor Bank, N.A. (Member FDIC), that natively handle ACH, wire, domestic and international transfers, and stablecoin transactions through a single API. Funds in the deposit account may be eligible for FDIC insurance; stablecoin holdings are not. Together, the two announcements represent end-to-end stablecoin payroll: enterprise stablecoin payouts on the send side, and FDIC-insured deposit accounts that natively accept stablecoin inflows on the receive side. [Sources: The Block, Infinite]

My Take:

A fully tokenized, fiat-interoperable payroll cycle is now demonstrable in production with two announcements published a day apart. Picture a Dasher: DoorDash settles their payout in stablecoin via Tempo. The Dasher routes that stablecoin into an Infinite-powered Erebor account. Erebor buys the stablecoin at a dollar, takes it onto its own balance sheet, and credits the Dasher with an FDIC-insured deposit. From there, the Dasher can send funds via traditional rails, hold them, or withdraw back out as stablecoin. The handoffs are invisible to the user. The economics are not — the Dasher avoids the 1.5% debit-push fee that has been the standard tax on instant gig payouts.

With stablecoins, DoorDash can avoid payout operations in 40 different countries. Stablecoin is the intermediate currency, but the use case is more about being cross-rail than cross-border. Local stablecoin-to-fiat liquidity is becoming a commodity service almost everywhere their workers live, which means that DoorDash can send value into almost any country using stablecoins and allow local players to deal with the complexity of managing local payouts. It’s more flexible, encourages local competition to drive down costs, and can scale quickly to support new payment methods.

The Erebor receive-side architecture is the model I keep advocating for banks. Erebor treats the stablecoin like a cash or FX deposit: when the dasher sends them stablecoin, the bank buys the stablecoin at $1, acquiring it as an asset on its balance sheet in its own name (not held FBO the dasher) and gives the dasher credit in their deposit account. The bank owns a stablecoin asset following the deposit, but it is under no obligation to continue owning that asset - it can sell those stablecoins for fiat which is redeployed in other banking products. With this model, the bank gets a modern payments rail that increases its deposit base with every stablecoin deposit, all without rebuilding its core.

There are now three U.S. banks that have shipped this: Cross River, SoFi, and Erebor. Someone once told me banks love to be first to be third. Now, banks can rush to be first to be fourth.

When AI Agents Can Pay, Who Collects the Toll?

American Express launched its Agentic Commerce Experiences (ACE) Developer Kit, which provides agent registration and verification, account enablement, intent analysis, payment credential verification, and digital cart context analysis, with Amex backing chargeback liability for errors caused by registered agents. Amex joins Mastercard and Visa which have both launched their own comparable offerings (Agent Pay and Trusted Agent Protocol, respectively). Separately, Coinbase-incubated x402 protocol unveiled Agentic.market, a marketplace where AI agents can autonomously discover and pay for online services across seven categories (Inference, Data, Media, Search, Social, Infrastructure, and Trading). The marketplace has already processed 2.3 million transactions and $50 million of volume across 118,000 buying agents, with a current run rate near 100,000 transactions per day. [Sources: The Block, American Banker]

My Take:

A year ago, agentic payments was a glimmer on the horizon. Now, we have base layer protocols for machine-to-machine payments, multiple active marketplaces, production scale volume, and integration into the frontier model companies whose own products already reach hundreds of millions of users. Despite their rush into the space, cards remain the less viable rail for agent-to-agent payments compared to stablecoins for the foreseeable future, and that’s where banks should plant their flag.

There are two distinct agent-payment use cases. The first is an agent buying things for itself: API calls, inference cycles, data feeds. With small ticket sizes and very high volumes, agents and their humans want the cheapest possible per-transaction cost. Stablecoins fit this perfectly with instant finality, non-intermediated value transfer. The second is an agent buying for its human: airline tickets and retail goods in our example. Higher ticket sizes, where chargeback rights, fraud protection, and rewards genuinely matter.

Agentic payments will grow fastest on stablecoin rails - not cards - for two reasons. First, cards face a significant hurdle in their economic model. None of the networks have announced any change to interchange treatment that would make the small ticket transactions economic. Visa’s Jack Forestell has gestured at the challenge, and I expect the networks will eventually experiment with ways to lower the minimum viable transaction size while attempting to protect the golden goose. That is going to take years to roll out at scale, and the incentive for participants to game any new schedule is enormous.

Second, barring an updated interchange model that works for small tickets, cards will remain the province of “human-scale” transactions. It will take time for users to get comfortable with autonomous purchasing at higher price points. Furthermore, the value of autonomous purchasing in human-scale transactions is a minor subset of the overall value provided by the agent. In a travel booking flow, the substantial work is in the research - from destination options, to weather averages, to flight times and beyond - not in the step of buying a ticket. Finally, as users move from agents-for-research to agents-for-buying, payments volume from a card issuer perspective won’t change much because these purchases are largely ones that the cardholder is already making.

What banks should be reading from this: the apps of today are about to be repackaged as agents that users fund per use. These can be consumer use cases or internal business use cases. The best user experiences are delivered today, and for the foreseeable future, with stablecoins, not cards. This is not about AI-native customers, it’s about serving anyone who is tech exposed (i.e. everyone to some degree). Therefore, banks should be looking at stablecoin payments as their next leg of growth. Even better, banks don’t need to create the demand for stablecoin payments in this strategy like they do for FedNow, RTP and other new rails. No, trillions of dollars are being deployed into the infrastructure, tools, and interfaces that will drive demand for stablecoin payments.

The Tokenized Deposit Use Case That Makes Sense

HSBC extended its Tokenized Deposit Service (TDS) to the United States, adding the U.S. dollar to a multi-currency platform that is already live in Hong Kong, Singapore, Luxembourg, and the U.K. The product enables corporate and institutional clients to move money between their own accounts across HSBC’s network 24/7. According to HSBC, the next step will be enabling clients to move tokenized deposits to other HSBC clients, with eventual interconnection to other private networks beyond that. HSBC joins Citi and JPMorgan among large banks that have deployed tokenized deposits. [Source: American Banker]

My Take:

Tokenized deposits still have a mystical allure — the idea that they will deliver everything stablecoins offer, with none of the perceived tradeoffs, and inside the bank perimeter. As banks actually build them, the assumptions that get waved past in conference panels start to harden into real blockers: centralized clearing, unlimited counterparty liability appetite, permissioned membership, and a near-total absence of cross-bank standardization. Stablecoins took five years to coalesce on a market structure and another five to get standardizing legislation. Tokenized deposits are at year zero on that curve.

The HSBC use case is a clean illustration of where tokenized deposits actually shine today, and where they don’t. Inside HSBC, a client in Hong Kong can move money to an affiliate at HSBC Singapore at any hour. Under the hood, value flows as a tokenized deposit recording that HSBC HK owes HSBC SG. The Singapore side credits the local account immediately and nets out via traditional rails later. This is a meaningful improvement on the status quo, and it might surprise some readers to learn that international banks routinely use SWIFT to move money between their own local entities — because banking is, structurally, a national activity. HSBC could have built this with bespoke inter-company accounting. They chose tokenized deposits because the rail, once built, is flexible: the same plumbing that moves an HSBC deposit token can easily extend to move deposit tokens from other banks, stablecoins, or any other kind of token.

The challenge facing 99.9% of banks is that they don’t have the foreign affiliate banks necessary to build a compelling first use case with deposit tokens. Some are looking at domestic tokenized deposit networks as a potential answer, though the few that exist are nascent and built on centralized technical architectures that complicate matters with issues of governance plus limitations on interoperability. The prospect of HSBC and other GSIBs opening their deposit networks to other banks is appealing on a volume basis, but reeks of another Early Warning System ownership dilemma for participants.

A different approach entirely is to start building with stablecoins, as Cross River, Lead, SoFi and Erebor have done, recognizing that interoperability can work from stablecoin to deposit token, or put differently, from where there is demonstrated growing demand today to where there may be a practical use case in the future.

Coupon Clippings

FinCEN is Rewriting the BSA/AML Rulebook

On April 7, FinCEN and the federal banking agencies proposed a sweeping overhaul of AML/CFT program requirements. The changes shift from process-driven, “check-the-box” compliance to risk-based, outcome-focused programs, with regulatory investigation and enforcement risk being redirected accordingly. The rules state that regulators should focus primarily on “significant or systemic failures” in already-established programs, rather than isolated or technical deficiencies. The framework reads like a description of the compliance practices that fintechs and crypto firms have already been building for years. For banks evaluating digital asset activity, the practical upshot is that examiners will now have a framework that maps cleanly onto the data blockchains actually produce like wallet attribution, multi-hop counterparty risk, Travel Rule data. Fine-grained data lets banks define more granular risk tiers than they ever could on traditional rails, where they only ever see the direct counterparty. This proposal is another regulatory tailwind for stablecoin services, and it deserves a comment letter from anyone planning to be in the business. [Sources: Perkins Coie, FDIC]

Thanks for reading. If you don’t hear from us for a bit, don’t worry—we’ll be back when there’s something important to share.

Omnia is a provider of stablecoin infrastructure for banks that want to capture growing demand for stablecoins. If you’re interested in learning more about us, please get in touch.

Getting value? Forward to a colleague and subscribe.