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

This week Cuy Sheffield is joined by:

🎙️ Listen to the latest episode of Tokenized here.

📷 Watch on YouTube.

The first stablecoin deployment inside a large company is usually its own money moving between its own entities. Eric Queathem said Velocity often leaves a first meeting with "maybe a dozen ideas" for stablecoins, then narrows to "largely internal first-party treasury movement." What gets companies experimenting, on his account, is a use case with low enough risk and clear enough guardrails.

Cuy Sheffield cited Hyundai's transfer of $20,000 from its US entity to its Mexican one through USDT, which took roughly seven minutes against a reported three to four hours through correspondent banking. Didier Lavallee called the amount minuscule for a company that size and read it as Hyundai's first transaction.

Canada extends the pattern into a developed market. Didier said corporate payments there still run T+2 or T+3, with no real-time business settlement nine years into the country's payments-modernisation programme. Roughly 80% of Canadian organisations trade with US counterparties, on his figures. Noah Levine saw room for stablecoins to "leapfrog the existing payment schemes."

The domestic gap is the part we'd watch. Correspondent banking is the usual target for stablecoins, but a company that cannot pay a supplier in its own country in real time has a problem stablecoins can address without crossing a border.

Velocity's $38mn Series A raise came with a claim about category. Eric pushed back on the orchestration label, which in acquiring means supplying technology without touching the money. Velocity is regulated, sits in the flow of funds, and offers services closer to a payments institution or bank, with one API covering sending, receiving, holding, growing balances and onchain credit.

Acquirer settlement gives that model a distribution route. Visa is opening infrastructure for existing clients to mint Open USD (OUSD), use their own or a managed wallet, and test transfers before launch. Settlement so far has run mostly through crypto-native card issuers, and extending it to acquirers pushes the flow out to merchants and their suppliers.

The funding route already exists inside the settlement relationship. A merchant can direct part of its card proceeds onchain by instructing its acquirer. “I want 5% of my settlement,” as Eric put it to Stripe, Adyen or Worldpay. The acquirer becomes a recurring source of onchain liquidity.

Seven-day-a-week settlement to acquirers remains a target rather than a production flow.

Didier and Eric differed on how far direct pairs can go, with their business models helping explain the split. Didier, as an issuer, backs the direct route. CAD-to-dollar conversion through a liquidity pool already beats the FX leg a financial institution would charge on cost and speed, he said, for transactions in the low thousands. His caveat was depth. "You don't have huge liquidity in these pools yet." Corporates moving hundreds of millions need more than the pools currently hold.

Eric was less convinced. Velocity is "probably less bullish on the ability to build enough liquidity to do those swaps outside of maybe a couple of core markets," he said, and expects primary liquidity to come from local currencies moving into OUSD, USDC and USDT rather than local-to-local pairs. Velocity's bet is breadth of venue, connecting to roughly 22 liquidity partners on Eric's count.

Iceland shows the asymmetry. Velocity found no one taking the other side of a USDC-for-local-currency trade, leaving a local bank and an offchain FX trade as the workaround.

That is the opening Noah described. He expects some banks to become the counterparty for these flows instead of ceding it to crypto exchanges, and in what he called "a crazy future," to provide liquidity to decentralised exchanges. Eric's version is blunter. In non-G5 corridors, whoever has the FX desk captures the conversion revenue.

Both positions can hold at once. Deep pairs may clear onchain while thinner corridors route across a bank balance sheet for years, which would make the FX desk a longer-lived asset than the direct-swap case assumes.

Cuy put current stablecoin supply at roughly $260 - $300 billion and contrasted it with forecasts of $3 trillion. He asked why a regulated institution would hold that in a stablecoin when it could hold a tokenized Treasury at the DTCC.

Eric agreed and split the assets by function. USDC works as a payment token, less so as a store of value. Tokenized Treasuries and money-market funds suit idle balances, while Eric placed AUSD and USDG in an emerging middle category offering more "Treasury-esque capabilities." If payment tokens are on-ramped, moved and off-ramped within minutes, he estimated supply might need to equal only "maybe 1/100 of total daily settled payment volume," even with global payments onchain. He flagged the figure as a rough estimate.

Didier did not fully accept the turnover argument. Tetra tracks supply alongside transaction volume, since Canada is roughly 6% of global FX flows on his numbers, and a yield-bearing Canadian token could attract foreign holders as allocations shift away from US Treasuries.

DTCC's 15 July production trades showed where the layers meet. JPMorgan tokenized Invesco QQQ shares, posted them as collateral against a margin call at CME Group and converted them back, which Cuy described as a central counterparty accepting tokenized assets as margin for the first time. BlackRock and Goldman Sachs also participated in the broader initiative, which involved nearly 40 companies and used Hyperledger Besu and Canton, according to Cuy.

Didier expects a "two or three tier system," with a stablecoin-led payments layer and a holding layer in tokenized deposits or Treasuries. Canada's six-bank market is likely to produce bank or consortium deposit tokens, and National Bank is a Tetra investor. Interoperability between the layers is the gap he named. "You don't go to JPMorgan and get a different dollar bill."

Visa and Artemis identified about $19 million of what the report classified as real x402 volume across 134 million transactions since May 2025, Cuy said. Roughly 4,000 wallets, the top 1% of buyers, drove more than 90%. Noah cautioned against reading the dollar figure as a verdict, since many transactions cost a cent or less.

Supply runs ahead of demand. Cuy cited more than 40,000 endpoints agents can buy from, against his own agents "sitting around waiting for me to tell them to do something." Assistant commerce takes instructions, while autonomous commerce would hand an agent an objective and let it discover tools and spend against it. Human time and creativity are the constraint.

B2B has the more legible economics. Eric put global AP and AR clerk labour costs at roughly $400 billion, then walked through how a CFO evaluates it: a $2 million cost base taken to $100,000, or $5 million of savings that a valuation multiple turns into as much as $500 million. Those figures were illustrative. Execution needs the network, which is why Eric expects Visa to solve the connectivity problem, with buyers and suppliers on both sides.

Cuy called agents plus stablecoins "risk squared," making AI-native companies adding stablecoins, or stablecoin-native companies testing agents, more plausible first movers than a traditional corporate taking both at once.

The next credible adoption signal is a payment network connecting enterprise buyers and suppliers, not another consumer-agent demo.

Thanks so much for reading the Tokenized Newsletter!

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

Disclaimers