The Stablecoin Banker, June 23

June 23, 2026

Three 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 or read back issues at thestablecoinbanker.com. If you're interested in stablecoin services for your bank, feel free to reach out to us at Omnia.

"Clients aren't necessarily beating down the door [for tokenized deposits]."
— Mark Monaco, Head of Global Payments, Bank of America (June 5, 2026)

In This Issue

• The biggest banks are building three tokenized deposit networks at once
• Mastercard, Visa, and Circle agree AI agents should pay in stablecoins
• Cross-border is about to become even better with stablecoins

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

Tokenized Deposits AND Stablecoins

JPMorgan, Bank of America, and Citi announced a tokenized-deposit network, to be operated by The Clearing House, that will let participating banks move deposits on-chain with 24/7 instant settlement, targeting launch in 2027. Days later, Early Warning Services - owned by JPMorgan, Bank of America and five others - named India as Zelle's first international corridor and unveiled ZelleUSD, a dollar-backed stablecoin intended for future cross-border markets. Visa also announced a tokenized-deposits feature for banks alongside its stablecoin-settlement and AI announcements. Sources [CoinDesk, American Banker, BusinessWire]

My Take:

In these announcements, we see the ongoing tension between tokenized deposits and stablecoin: when each type of token makes sense, the network effect challenge, and the urgency that banks are applying to reconciling the tension.

On the point of urgency, the banks have now birthed three major tokenized deposit projects, with substantial overlapping membership and ownership. In March I wrote about Cari, the tokenized-deposit network from five regionals, and argued it was really a tokenized version of CHIPS/RTP's bank-to-bank payments with net settlement. FIS announced their version of the same strategy last month as Project Keystone. Three banks have now signed on to both Cari and Keystone (Huntington, M&T, and KeyBank), and all three are also owners of The Clearing House, which is busy standing up the third network. The overlap is comical, but probably reflective of a degree of urgency felt by the banks as the yield debate on stablecoin appears to be closing, GENIUS act final rules are coming due, and adoption throughout mainstream finance continues unabated.

It also points to the structural coordination challenge facing the banks. That stablecoins have reached $300B of market cap and trillions of volume is not just a circumstantial outcome of open blockchain infrastructure, but a direct result of decentralization. Decentralization aligned incentives of early adopters to use common blockchains, token standards, and interfaces. Banks have historically struggled to coordinate through centralized structures over time, whether that's NACHA which became sclerotic or Visa and Mastercard which turned into a love-hate relationship for banks, or TCH and EWS which represents the interests of only the largest banks. Will the haphazard approach we're witnessing yield a new solution, or more of the same? I have no idea, but the effort is probably still worth it.

Tokenized deposit networks have great potential for banks, provided they can find viable implementation and application. They keep the credit, regulatory, and accounting treatment of a deposit. They also enable net settlement, which brings capital efficiency that's uniquely valuable for a bank, which can earn high rates of returns on its assets provided it can reasonably expect it won't need cash on hand for withdrawals. This is where the roadmap for Cari, Keystone, and TCH is most important. They need to graduate beyond tokenized RTP, and if they can do so, by bringing more use cases onto the network than simply interbank transfers, they can make banks more capital efficient than ever before by expanding the holder base of their deposits. This will require figuring out how to extend membership to non-banks, which is non-trivial from a compliance, risk, and operational perspective (for further perspective on this challenge, read the Deposit Token section of our new paper, Banking Stablecoins)

Until these challenges are figured out - which I put on a multi-year timeframe at best - stablecoins will continue to provide utility outside the private interbank networks. But, whether stablecoins are truly required, even in cases like cross-border, is an open question for companies at scale. That question is what makes the Zelle announcement most illuminating. EWS simultaneously announced a stablecoin alongside their first cross-border corridor, which uses …. traditional rails. Why? First, for a single corridor, a large provider like EWS doesn't need a stablecoin. They simply need an Indian bank willing to hold a nostro account for EWS, a way to make third-party payouts directly or through a local MSB, and a net-settlement flow over SWIFT. Second, India also happens to be openly hostile to dollar stablecoins.

While EWS certainly can replicate this strategy in other areas, doing so quickly is where stablecoins make the difference. Connecting to hundreds of other cross-border corridors within the banking system, plus the hundreds or thousands of non-bank wallets in existence is practically impossible. Instead, any property that supports stablecoins is much more easily reached by EWS using stablecoins. In the last newsletter, I wrote about how stablecoin swaps will become widely adopted to allow companies like EWS to both issue their own stablecoin and maintain interoperability with properties that don't accept it.

Here's the takeaway for a community or regional bank: the industry has quietly resolved the question of tokenized deposit or stablecoin. For banks the answer is "yes". You will need to support both. Visa has thrown their weight behind stablecoins, and now Zelle is joining the party. Fortunately, this is not a Betamax vs VHS dilemma, where choosing the wrong technology too early disadvantages you. The principles and infrastructure that apply to stablecoins today will carry over to tokenized deposits tomorrow. Enabling stablecoins today is a bet on tokenization that will serve you as soon as tokenized deposits are proven.

Dark Fiber of Payments

Announcements for agentic payments continued their rapid pace. Mastercard launched Agent Pay for Machines, which lets AI agents transact autonomously across cards, bank accounts, and stablecoins, with more than 30 partners including Coinbase, Stripe, and Solana. Visa unveiled an agentic stack that includes Agent Score for merchant AI-readiness evaluation; an agentic directory to intermediate trust between merchants and agents; a large transaction model for fraud detection; and an OpenAI partnership. Circle released its Agent Stack, comprising Agent Wallets, nanopayments, and an Agent Marketplace built around USDC. Across all three, stablecoins sat at the center of machine-initiated commerce. Sources [Mastercard, Visa, Circle]

My Take:

This week I noticed a narrative shift around agentic payments. Both Visa and Mastercard have moved away from canonical retail agentic payments use cases like travel booking, and are leaning much more into business use cases. This probably follows the overall AI narrative shift in the same direction, alongside impressive results and valuations from Anthropic, widely viewed as the current leader in enterprise AI use cases. Mastercard, for example, outlined a scenario of a small business building a website with an AI designer, which could procure assets, hosting, SEO, etc. within the design and deployment flow. Visa's products focus on ensuring merchant sites are "agent ready" and closing the trust gap between agents and merchants (i.e. can an agent trust the merchant and vice-versa).

That shift is particularly important for banks in the context of the card networks' push for stablecoin settlement. I've previously written about how stablecoin-backed cards, which settle to the network in stablecoins, creates a flow of funds model and prefunded risk model that obviates the need for a bank. Agentic payments make this end-run even more viable because they don't require the networks figuring out how to empower more non-bank card issuers (which would certainly enrage their bank customers). A business user simply funds a stablecoin wallet, which is as simple as issuing a single command to an AI chatbot. Tools for ERP integration, project-level cost accounting, and even company-in-the-middle transaction approvals are all viable, if not already in existence, and make stablecoin payments even more appealing to business users.

It's clear why Visa and Mastercard are rushing into this space: they need to position themselves in the center of any financial transaction network. But, how they will make money is a very foggy picture: Mastercard's CPO Jorn Lambert recently explained the company doesn't expect agentic payments to be a major revenue driver in the near term, and Visa's CFO Chris Suh echoed that sentiment. This is a land grab for position, not a business plan. However, they are starting to assemble a constellation of services that all seem logically relevant to a robust agentic + stablecoin payments system.

Admittedly, this is all still extremely early. Various data sources place true agentic AI adoption at a microscopically small level. Actual agent-initiated payment volume is essentially zero. But "I don't know the timing" is not the same as "I can ignore it." On a recent episode of Invest Like The Best, Alex Sacerdote, a tech investor, discussed the dynamics of S-curves. He illustrates how AI is really at the early innings of adoption, and the thing that constantly surprises most people is how dramatically the slope of the S-curve changes. Software adoption curves are particularly steep because the solution is distributed over existing pervasive communications and hardware surfaces. Stablecoin is setting up for a similar story: the rails and base layer infrastructure like wallets and on-ramps are already built. Visa, Mastercard, Circle, Google, and all the others stumbling over themselves to build the transactional framework for agentic payments is creating the "dark fiber" of stablecoins ready to light up.

When machine-initiated payments do scale, they will run on stablecoin rails. All three of these announcements agree on this point. A bank that can't supply stablecoins can't be the funding source for its customers' agents. It's early, so don't try to predict the timeline; but deploy the basic wallet and stablecoin functionality now. Building that capability while volume is still zero costs very little and buys the option to be at the center of those flows when they arrive.

Stablecoin Cross-Border Set to Scale

Circle and Nium connected USDC settlement to Nium's payout rail across more than 190 countries through a single API, letting institutions send dollars and have them arrive in local currency without assembling a patchwork of correspondent-banking relationships. Separately, the Federal Reserve proposed amending Regulation J to let FedNow participants route the domestic leg of a cross-border payment, opening the US instant rail to international flows; the comment period closed June 9. Abroad, Japan's three largest banks — MUFG, Mizuho, and SMBC — announced plans to jointly issue a yen-denominated stablecoin under a trust structure by March 2027, while Qivalis, a consortium of 37 European banks, remains on track for a MiCA-compliant euro stablecoin in the second half of 2026. Sources [American Banker, ABA Banking Journal, CoinDesk]

My Take:

Cross-border is the use case stablecoins have already won. This week three different players moved to make it faster.

It was easy to interpret the Fed headline as a threat to stablecoin cross-border, but its proposal sounds bigger than it is. Today, FedNow can only move money between a sender's bank and a recipient's bank. By contrast, most cross-border flows must route through one or more intermediate correspondent banks. That daisy chain of banks is permitted on Fedwire but barred under FedNow's rules. The proposed change would simply permit the US leg of a SWIFT payment, from the US correspondent to the recipient's bank, use FedNow. It's useful, but incremental because it does nothing for the overseas legs, and it doesn't address the operational complexity that obscures and slows down the payment flow.

The Circle–Nium tie-up is more concrete: one API now does what a web of correspondent relationships used to, settling USDC out to local currency across 190-plus countries. There are a number of stablecoin payment "networks" out there: Circle Payments Network, Conduit, Bridge, and so on. They all suffer from limited geographic reach and capabilities. The potential of the Circle-Nium partnership is to break through these barriers.

The more structurally relevant news came from the Japanese megabanks and European Qivalis consortium building trusted yen and euro stablecoins. Today, essentially all fiat-to-fiat stablecoin payments run on dollar coins, and the FX conversion happens at the on/off-ramp. That fragments liquidity: each ramp provider has to supply both the liquidity (swapping one asset for another) and the FX (dollars into local currency), which makes those providers specialized, expensive, and hard to find, and it caps how large a payment they can clear.

The payoff is simple: separate the currency swap from the crypto on/off-ramp, and each side gets deeper, cheaper, and able to clear far larger payments. Well-regulated non-dollar stablecoins issued by large institutions break that bottleneck by separating FX from ramping. Ramping into a yen or euro coin in the respective country becomes a matter of using local capital, denominated entirely in the local currency, which is far more abundant and priced at a spread over the risk-free rate with no embedded FX cost. That alone lets you on- and off-ramp much larger payments. Then, with both sides of the FX trade in tokenized form, FX liquidity can centralize in a limited number of large and trusted on-chain venues, spreads tighten, and volume grows. The issuers are incentivized to support both legs with their own capital because they earn yield on the outstanding stablecoin reserves.

Here's the part bankers talk themselves out of. Most community and regional banks wave off cross-border because they don't do much of it. But look at who sends those payments: they're disproportionately commercial customers, the bank's highest-value relationships. While international payments may be rare, each bad experience is one more reason for that commercial customer to take a look at a Mercury or a JPMorgan that offers seamless cross-border and a cleaner, more modern banking experience overall. This is not really a payments story; it's a deposit-retention story that is told by looking not just at your own capability but at how your capability compares to the competing offers landing in your client's inbox every day.

Coupon Clippings

Choke Point Is Dead. What's Your Excuse Now?

The OCC, FDIC, and Federal Reserve put into effect a final rule and reissued fifteen pieces of guidance that strip every reference to "reputation risk" and bar examiners from acting against a bank for offering a lawful product. Industry observers called it the formal burial of Operation Choke Point 2.0. The same day, the OCC published Interpretive Letter 1192, confirming that national banks and national trust banks doing digital-asset work aren't subject to state money-transmitter licensing or state visitorial authority over those activities, expanding the comparative advantage of a national charter to state charters. The fear of regulatory reaction among bankers considering digital asset products has been removed in writing, five weeks before the GENIUS Act's July 18 final-rule deadline. Sources [FDIC, OCC IL 1192]

Fiserv's FIUSD Lands in July. Is It a Stablecoin or a Deposit Token?

Fiserv confirmed that FIUSD launches in July, with Bank of North Dakota as an early pilot. FIUSD is positioned as infrastructure for community banks and credit unions to offer tokenized-dollar balances and on-chain settlement inside the regulated system. While the potential for the product to come embedded in a core used by one third of the US banking industry is impressive, its identity is muddy. Fiserv talks about FIUSD as a stablecoin, yet the first use case runs through Bank of North Dakota, which is a pure correspondent bank for the state's community banks. This is an interbank job befitting a deposit token, not a stablecoin. There's a hybrid path, a stablecoin backed entirely by bank deposits, but GENIUS's coming limits on how much of a coin's reserves can sit at any single bank make that structure look Frankenstein: correspondent banks passing around a coin stitched together from deposits at many banks, instead of BND simply issuing a deposit token. It's hard to see how that makes sense in the very first use case they've chosen to show, so I'll be watching closely. [Full Story]

Digital Asset Raises $355M for Canton — and the Public-vs-Private Question Is the Wrong One

Digital Asset raised $355 million, led by a16z with HSBC, BNP Paribas, CME, and Apollo participating, to scale the privacy-preserving Canton Network as on-chain infrastructure for capital markets. I've argued consistently that private, permissioned chains are the weaker path because they struggle to attract the users and liquidity that make a network valuable. For payments, I'll stand by that, but Canton is an institutional chain for institutional jobs like securities settlement, repo, and financing where the participants only need to transact with each other, strictly require privacy, and don't need to interact with a wider ecosystem. This round is not a verdict on the public-chain thesis. The likelier outcome is that banks live in a multi-chain world: part of the balance sheet tokenizes and trades on a permissioned venue like Canton for institutional liquidity, while the part that touches businesses and consumers for payments runs on public chains. Treating "public versus private" as an either-or choice misses that a bank will run on both, for different reasons. [Full Story]

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.