Six Canadian Banks Want Tokenized Deposits, Not a Coin
The stablecoin headlines this week are loud. The more useful sentence is quieter, and it is about tokenized deposits, not a ticker. PYMNTS on Sept. 24 looked at a few days of payments news and argued that stablecoins have changed the rails without capturing the economics. The fight, in that framing, is over the customer, the wallet, and the payment flow. The token is the prop.
Two facts in that piece are new enough to stand on their own. Binance invested $100 million in Circle and signed a five-year agreement to expand USDC. And on Tuesday, Sept. 22, Canada’s six largest banks said they are exploring a network for tokenized Canadian-dollar deposits that can move between them. BMO, CIBC, National Bank, RBC, Scotiabank, and TD. PYMNTS attributes the Canadian item to that day’s press release. I have not read the release PDF. I am using the Sept. 24 account of it.
We already wrote SoFi settling its Mastercard book in SoFiUSD. PYMNTS repeats the figure, a card program expected to process more than $25 billion a year, with transactions already live. I am not walking that plumbing again. SoFi is the example of a bank burying a stablecoin under a card the customer already holds. Canada is a different move. Those six banks are not announcing a coin you can withdraw to a personal wallet. They are talking about deposits they already owe you, represented on a ledger that can settle after the branch closes.
Tokenized deposits are the banks’ copy of the rail
PYMNTS draws the line cleanly enough that I will keep their distinction rather than invent a better one. A tokenized deposit is a regulated bank deposit held on blockchain infrastructure. The economic relationship does not change. You are still a depositor. The bank still owes you the dollar, or in this case the Canadian dollar. What changes is the settlement mechanism. The token is a record of a deposit, not a new liability issued by a nonbank.
That is why this is not a sequel to the 21-bank dollar stablecoin plan. That story was a consortium talking about its own coin, aimed at mid-2027. A coin has an issuer, reserves, and a redemption promise that can sit outside the deposit franchise. A tokenized deposit is the franchise, with a faster back office. If the Canadian exploration becomes a live network, the thing moving between BMO and TD is still a bank deposit. If it stalls at “exploring,” it is a press release, and I will have over-read a Tuesday.
PYMNTS’s sharper claim is about who keeps the economics. Stablecoins got fast, and they got programmable. Banks noticed. Tokenized deposits are the attempt to copy the speed and the programmability without handing the funding economics to a stablecoin company. Funding economics is the dull phrase for a real fight: whose balance sheet the money sits on, who earns the float, who the regulator calls when the ledger and the cash do not match. A USDC in a brokerage app and a tokenized TD deposit are both “digital dollars” in a headline. Only one of them is a deposit at TD.
We have a prior data point that this is not a thought experiment. HSBC and Swift already moved a tokenized deposit in production. Canada’s announcement is earlier in the pipeline than that. “Exploring” is not “live.” The reason to write it this week is the cluster, not the adjective. A G7 banking system saying it will study interbank tokenized deposits, in the same week a card network is settling a bank’s own book on a stablecoin, is the industry picking a lane. The lane is: keep the customer relationship, change the settlement clock.
I do not know the governance. Six banks exploring a shared system have to answer whose node is canonical, what happens when one bank’s tokenized deposit needs to become another’s on a Sunday, and which regulator gets the incident report. PYMNTS does not have those answers, because a press release about an exploration does not have them. The absence is the story as much as the names. A live rail would name an operator. This one names a committee.
The $100 million is distribution, not a new rail
Binance’s $100 million into Circle, with a five-year agreement to expand USDC, is the other half of PYMNTS’s week. I do not have the term sheet. I have the two facts the Sept. 24 piece states: the investment size, and the length of the commercial agreement. Do not promote that into a valuation, a stake percentage, or a promise that USDC volume on Binance will rise by any particular amount. None of that is in the account I am using.
Read it next to the Canadian release and the contrast is the point. Circle and Binance are spending money and years to push a stablecoin through an exchange’s distribution. The Canadian banks are trying to get the settlement properties of that rail without issuing the stablecoin. One side wants the token in more wallets. The other side wants the wallet to remain a bank account, with the token as plumbing the customer may never see.
PYMNTS also says the European Central Bank and the EU’s national central banks pushed for changes to how stablecoin reserves interact with commercial banks. The extract does not quote the proposal. I will not invent a reserve ratio or a ban. The reportable fact is the push, in the same few days as the Binance check and the Canadian exploration. Central banks are arguing about how a stablecoin’s cash touches the banking system, while banks are building a version of the rail that never leaves the banking system. Those are the same argument from opposite ends.
Thredd, an issuer processor, expanded its platform to include stablecoin-powered money movement. That is one sentence in the PYMNTS piece, and it should stay one sentence here. A processor adding a rail is how this stuff becomes boring infrastructure. It is not, by itself, a volume number.
SoFi remains the cleanest live example of “the customer never holds the coin,” and it is already written. The new information is that the same week produced a distribution deal for USDC and a deposit-token exploration by six Canadian banks. If you only track the coin announcements, you will miss the copy. The copy is the product the banks would rather you use.
Tron’s week is the volume they are staring at
There is a number attached to the rail the banks are trying to imitate, and it is not a Canadian number. Josh Olszewicz, in CoinDesk’s Crypto Long & Short, published Sept. 23 and updated Sept. 28, writes about Tron as a settlement network for stablecoins. The column carries the usual disclaimer. He is a portfolio manager and head of trading at Canary Capital. The views are his, not CoinDesk’s. I am taking the network facts he attributes to Tron’s explorer. I am not taking a view on TRX, and this is not a suggestion to hold any token. A volume statistic is not a trade.
The facts, as that column states them: Tron uses delegated proof of stake. TRX holders stake for voting power and elect 27 Super Representatives, the validators that produce blocks. Weekly transaction counts have climbed toward record highs, recently approaching 100 million transactions a week. The average on-chain fee has fallen to around seven cents, which he calls a multiyear low. Weekly active addresses are climbing toward records. Stablecoin transfer volume on the network has been running at roughly $150 billion to $190 billion a week.
I have not queried the explorer myself this morning. If you are the person who has to brief a risk committee, open the explorer and check the week, do not brief from a columnist’s chart description. The range is wide enough to matter. One hundred fifty billion and one hundred ninety billion are not the same week. “Roughly” is doing work. Even the low end is a settlement number that makes a $25 billion annual card book look like a pilot, which is a comparison of a stock of annual volume to a flow of weekly transfers, so do not divide them and call it a market share. They are different units. The point of putting them in the same article is scale, not a ratio.
Twenty-seven elected validators is the part a bank lawyer will circle. Fast confirmation and a seven-cent fee come from a small, elected set of block producers, not from a chain where anyone with hardware can write the next block. Olszewicz describes that concentration as the reason Tron can keep overhead low. A Canadian interbank network would concentrate production too, just inside institutions that already have a regulator and a deposit license. The technical shape rhymes. The legal object does not. Tron’s stablecoin transfers are transfers of tokens issued by stablecoin companies. A tokenized TD deposit is TD’s liability. Copying the clock does not copy the issuer.
I am leaving his comments on legislation alone. A column by a trading-desk PM about what a proposed market-structure bill might do to a token is the edge of investment commentary, and we are not crossing it. The explorer figures are enough. They say a public chain is already settling stablecoin volume in the hundreds of billions a week, at a fee measured in cents, on a validator set you can count on two hands. That is the benchmark the Tuesday press release is implicitly answering, whether or not the six banks named Tron.
What is live, and what is a meeting
Sort the week by what a customer can touch.
SoFi’s card settlement is live, on a book the bank says will clear more than $25 billion a year. We covered the merchant side: the merchant still gets cash. Binance’s $100 million and the five-year USDC agreement are a signed commercial relationship, as PYMNTS reports it. They are not a new chain. Thredd adding stablecoin movement to an issuer platform is a vendor feature. The ECB push is a policy argument whose text I do not have. Canada’s six banks are exploring. Exploring is a meeting, plus a press release, plus a list of logos.
The mistake is to rank them by how crypto they sound. The live card settlement and the signed USDC distribution deal are further along than the Canadian exploration. The Canadian exploration is the one that, if it becomes a network, keeps the deposit economics inside the banks. PYMNTS’s thesis is that this is how the system absorbs the rail. I think that thesis fits the facts they assembled. I also think “exploring” can die in a working group, and a $100 million check can be a partnership announcement that does not move volume. Both are allowed. Neither is a reason to treat a token price as the story.
If you run payments at one of those six banks, the question to take into the exploration is the one PYMNTS already answered in the negative for stablecoin companies. Who keeps the float, and who shows up on the incident call? If the answer is still the bank, the tokenized deposit did its job. If the answer slides toward a technology vendor or a stablecoin issuer, you have built SoFi’s plumbing without SoFi’s charter, or Circle’s distribution without Circle’s token, and you should notice before the pilot has a name.
I am not telling you to buy USDC, TRX, or a Canadian bank. I am telling you the week’s blockchain news, read as payments news, is a fight over the liability. The rail is being copied. The coin is what the copiers would like to leave out.
