If you're reading this and still haven't signed up, click the subscribe button below!

Pssst. We also have a podcast… find that here on your favourite podcast player or here on YouTube 🙏

Join Tokenized and Modern Treasury for Dinner in Miami during Sibos.

Next Tuesday, September 29, we’re hosting an intimate dinner with Modern Treasury, bringing together founders and operators shaping the next generation of financial products.

Space is limited, so RSVP here if you’d like to join us.

https://go.moderntreasury.com/4rcxLPW

When SoFi launched SoFiUSD in December, we said it would become a platform business if two or three regional banks or major fintechs signed on to its rails. The first large live flow turned out to be SoFi's own card book.

The bank has switched on SoFiUSD settlement across its Mastercard debit and credit program and is migrating the full program, expected to process more than $25 billion a year, onto the rail.

That sharpens the question we left open in May: why would anyone outside SoFi hold SoFiUSD instead of USDC?

SoFiUSD doesn’t need people to choose it first. The settlement flow already exists with Mastercard.

Card settlement puts SoFiUSD against an obligation SoFi already has on the Mastercard network. The stablecoin does not need to create new consumer demand. It needs to become an accepted settlement asset for a flow that already exists.   

The deposit stays inside SoFi. The stablecoin does the traveling.

This is why SoFi runs two rails. In May we covered the dual rail as a fix for 24/7 payments and the par problem. The tokenized deposit earns interest and carries FDIC insurance, but only SoFi members can hold it and it cannot be sent externally. SoFiUSD is the leg that can leave the building, and Kraken now gives a non-SoFi holder somewhere to sell it, though redemption at par stays with approved SoFi customers.

Other banks are solving the same exit problem differently. JPMorgan, Citi, Bank of America, Wells Fargo and others are building interoperable tokenized deposits through a shared network operated by The Clearing House (TCH), targeted for H1 2027. Swift is working on a shared ledger that connects banks' existing tokenized deposit systems. SoFi's version is already settling card volume.

SoFi pays Mastercard in money it issues itself.

The first treasury benefit is the weekend. An issuer settling on traditional rails has to pre-position cash against card spending that continues while settlement windows are closed. Seven-day settlement shrinks that buffer, the same problem we covered when Mastercard announced always-on settlement in June. The bigger difference is who funds the bill. Visa says some early-stage stablecoin-linked programs are constrained less by demand or network infrastructure than by working capital. Rain is a good example, having financed ~$2 billion of Visa settlement through a Credit Coop revolving facility since August 2023, per Visa.

SoFi still has the funding requirement. It internalizes it. The bank funds settlement from its own balance sheet, with no third-party stablecoin to source and no separate credit facility to borrow one. A fintech card program may need that external financing layer; SoFi keeps it in-house. That funding stack is what the charter buys in production.

SoFi also controls the economics of the settlement asset. An issuer settling in USDC holds a third party's coin while Circle captures the reserve income. SoFiUSD is issued against SoFi's own reserves, so SoFi decides how those economics are shared. The trade-off is velocity: the shorter settlement balances remain outstanding, the less float there is to earn on. Weekends and holidays are where that advantage should matter most.

OpenUSD, which passes reserve earnings to partners after a management fee, is built for issuers without that control. SoFi already has it.

Watch for a weekend bump in SoFiUSD supply.

If settlement balances are turned back into dollars quickly, supply stays small even as billions a year pass through. If they sit over weekends, supply should rise on Friday and fall on Monday, once card flows are large enough to stand out from Kraken trading.

The next test sits with Galileo’s clients.

If a Galileo issuer switches within the next two quarters, SoFi has a settlement platform.

If none does, SoFi has built a very good way to settle its own Mastercard bill.

The ECB spent 2024 testing DLT settlement with 64 participants. Two years later, Pontes is live. An initial cohort including Deutsche Bank and Santander can now settle tokenized trades against central bank money through TARGET. That lands awkwardly for the 37 financial institutions behind Qivalis, a euro stablecoin pitched partly as onchain settlement infrastructure and still waiting on its Dutch licence. If those banks already have risk-free settlement, what is a bank-led euro stablecoin for?

Pontes takes part of the case we made for Qivalis in December. We framed Qivalis partly as interbank treasury infrastructure that could shrink pre-funded nostro balances, and the securities-settlement slice of that case now sits with the ECB.

For a tokenized bond trade between Deutsche Bank and Santander on a Pontes-connected platform, there is little reason to introduce an EMI claim when central bank money is already available for settlement. Qivalis still has a role in cross-border and onchain payment flows, particularly where the money needs to circulate beyond institutions with access to Pontes and TARGET.

Qivalis’s stronger case sits beyond the settlement perimeter.

A Pontes transaction finishes inside the wholesale perimeter. Qivalis can keep moving across corporate treasury, exchanges and fintech products, with a public-chain design that extends into smart contracts and collateral markets. The hours advantage disappears if Pontes delivers 24/7 service by mid-2028. The holder-base advantage does not. In that lane Qivalis competes with Circle's EURC and the dollar stablecoins that already dominate onchain markets.

Pontes also leaves the legal claim unchanged. A Qivalis token remains an e-money claim on Qivalis rather than central bank money, whatever rails its member banks use for their own flows.

The digital euro sits on a different axis. It is public money built for retail payments, while Qivalis is private money built to circulate onchain. Digital euros replace commercial bank deposits with a central bank liability, helping explain the proposed holding limits and zero remuneration. Qivalis also pulls money from customer deposits, but part of its reserve backing remains inside commercial banking.

Sustained euro liquidity onchain would weaken the argument that Europe needs public digital money to answer dollar stablecoins. It does less to the separate case around retail payment sovereignty, resilience and dependence on non-European payment infrastructure.

The Commission proposed the digital euro in June 2023, and trilogues are now underway. If Qivalis secures authorisation and Pontes grows beyond its launch cohort, Europe could have wholesale central-bank-money settlement and a private onchain euro running well before potential digital euro issuance in 2029.

The project Europe has debated longest may be the last to arrive.

Canada has been waiting years for instant payments. Real-Time Rail (RTR), first promised for 2019, starts rolling out this quarter. Weeks before it does, the country's six largest banks said they want to build a shared tokenized deposit network.

That raises an awkward question. Once Canadian banks can move money instantly, 24/7, what is the token for?

Real-Time Rail takes away the easy sales pitch for tokenized deposits.

Once it is live, Canadian banks won't need a blockchain to make an everyday payment instant. RTR is designed to do that for anything up to C$100,000.

So the token has to earn its place somewhere else.

The most promising place is where money meets assets. Project Samara showed how that can work: a bond issued and traded on DLT, with the securities and cash legs connected so they could settle together. But Samara used central bank money (W-CAD), which only banks with Bank of Canada accounts can hold. Tokenized deposits could eventually extend that kind of settlement to the commercial bank money used by corporate and institutional clients. Larger treasury flows and programmable payments could follow, though phase one is only about moving deposits between banks.

Canada is well placed to try. The Big Six hold ~ 94% of deposits, according to DBRS Morningstar's 2024 outlook, so six banks agreeing on one model gets most of the way to a national system. Two of them, BMO and TD, also sit in The Clearing House's tokenized deposit network in the US, which gives them a view of both approaches.

National Bank is keeping its options open. It is in the deposit consortium and is also an investor in Tetra, the CADD issuer.

Then there is the question of who gets in.

Payments Canada has opened membership to PSPs and credit unions, with Wise and KOHO among the newcomers now eligible to apply for RTR participation. Canada is opening the door to its national payment infrastructure. Its tokenized deposit project starts with six banks.

If smaller institutions can't join directly, they may end up reaching tokenized markets through one of the Big Six, recreating some of the dependency Canada's payment reforms are trying to reduce.

Thanks so much for reading the Tokenized Newsletter!

Please share this edition or share it with your colleagues if you enjoyed it!

Disclaimers