Wall Street has found a use for blockchain that involves neither memecoins nor replacing the banking system. It’s now helping institutions borrow billions between breakfast and lunch, usually because they have government bonds but not enough cash when a payment comes due.
Getting a bill paid on time might be the technology’s least glamorous selling point, but banks have good reasons to care.
Broadridge, a financial technology company with little of crypto’s celebrity, says its distributed-ledger repo platform processed an average of $359 billion a day in September, or $7.5 trillion across the month.
Its system helps institutions finance securities, move collateral, and sometimes borrow for a few hours instead of the usual overnight repo borrowing. Not every transaction it processes is an intraday loan, and the numbers measure activity rather than loans outstanding, but the scale is nothing short of extraordinary for a tradfi blockchain application.
Banks can get the money they need at the wrong hour. One trading desk might own billions in US Treasury bonds, owe another institution cash at 9 a.m., and expect an equally large receipt at 11. Those bonds are valuable, but they can’t settle the bill without selling them or using them as collateral for a loan.
The world’s most expensive two-hour cash problem
Repo is short for repurchase agreement, and it’s a fairly straightforward borrowing deal. One institution temporarily sells securities, often government bonds, to another in exchange for cash, with an agreement to buy them back later at a slightly higher price that reflects the cost of borrowing.
Economically, it works much like a secured loan: the borrower gets money, the lender holds securities as collateral, and the parties agree on when the arrangement ends.
Money market funds, banks, and other institutions with cash available can be on the lending side, while securities dealers, banks, and hedge funds use repo to finance bond inventories or trading positions.
It keeps people from having to sell long-term investments every time their short-term obligations come due, and it’s one reason the US Treasury market can support so much trading. The Office of Financial Research estimated that more than $8 trillion in Treasury-backed repo was outstanding in August 2025, much of it routinely refinanced as loans mature.
Overnight repo is straightforward: borrow against bonds today, then reverse the deal the next business day and pay the financing cost.
Take a hypothetical dealer that must send $1 billion at 9 a.m. and expects $1 billion back at 11. If it borrows overnight at an annualized rate of 5%, one full day of interest at the usual 360-day money-market convention works out to about $138,889.
Pricing two hours at the same annualized rate on a simple proportional basis would cost roughly $11,574, leaving a theoretical difference of $127,315 on a loan that never needed to last the entire day.
But we shouldn’t mistake that for a quote Broadridge is offering. Intraday deals have their own rates, and an overnight repo isn’t always billed as if a borrower could slice the contract into 12 two-hour pieces.
The example just explains why institutions bother: on very large sums, the price of borrowing for hours they don’t need can turn into real money, even when the interest rate looks tiny.
Borrowing for only two hours requires the parties to establish which securities are available, who controls them, whether the money has arrived, and when the collateral can be returned.
Everyone involved needs matching records, so while the loan might be short, the process and communication needed to arrange and close it can be pretty long.
Tokenized repo is blockchain’s least glamorous killer app
Broadridge’s distributed ledger repo system, usually shortened to DLR, tries to shorten that chain further by giving participants a shared record of the transaction.
Its explanation of how the platform operates says underlying securities are held within established custody arrangements and represented by digital tokens that let the system transfer rights to the collateral.
Contract instructions coordinate the handover of those rights with confirmation of the cash payment, then reverse the process when the borrowing is repaid.
The distinction from the crypto most people know is one of the most important things here. Banks aren’t necessarily turning their Treasury portfolios into freely circulating coins, and Broadridge’s live arrangement doesn’t require the loan itself to be paid in a new blockchain currency.
The cash leg runs through conventional payment systems, while the ledger tracks the securities and the obligations attached to the deal. Smart contracts (the software that carries out agreed instructions) help the institutions coordinate those moving parts so that one side isn’t left waiting for the other to update its records.
Someone used to seeing blockchains pitched as replacements for banks might find this anticlimactic. We’re just using the technology to make banks less cumbersome to use for institutions, with old-fashioned dollars moving through old-fashioned payment channels.
The improvement (for institutions) comes from knowing exactly where collateral is and who can use it, especially when a security needs to move between counterparties or across borders. Broadridge expanded DLR to securities from G7 markets in September, which makes that coordination more useful than a system confined to one country’s bonds.
Neither the ledger nor the tokens create another pile of cash for Wall Street. September’s $359 billion daily number includes platform-reported repo processing across supported products and collateral types.
Bonds can be used in successive financing transactions, and activity totals don’t show how much collateral appears repeatedly. The New York Fed’s overview of the repo market puts transactions underlying the SOFR overnight funding benchmark at roughly $3 trillion daily, but the two statistics cover different business and can’t be divided to produce a reliable market share.
September’s enormous total also looks different next to Broadridge’s other 2026 announcements. Its reported daily average was $365 billion in January, $362 billion in February and May, and $359 billion in September.
That’s a decline of about 1.6% from January to September, even though its press releases described large increases compared with the same months in 2025. The company could have added customers or handled different securities during that time, but the four public observations don’t show a steady climb in daily volume throughout 2026.
Source: Broadridge’s announcements for January, February, May and September. These are selected published months, not a complete series; different numbers of business days affect monthly totals.
Where the savings actually come from
In a Broadridge-published white paper, Société Générale’s Michael Walsh described eliminating spreadsheet work and almost $1 million in annual operating costs.
He estimated that daylight-overdraft charges, incurred when banks temporarily need credit to complete payments, had fallen to zero on four out of five days.
Anyone who has reconciled spreadsheets kept by different departments can appreciate the appeal, although these entries are a little more consequential than an expense claim.
Delays in releasing collateral tie up resources, while borrowing overnight to cover a two-hour gap costs money. Banks can spend less on both if transfers are better coordinated.
Broadridge and consultancy Finadium modeled another benefit: if 15% of a bank’s intraday funding used DLR, the bank’s required liquidity buffer could fall by an estimated 8% to 17% under their assumptions.
Liquidity buffers are cash or other readily available resources a bank keeps to meet payments without having to hunt for an emergency loan. Smaller buffers can make ordinary operations cheaper, provided the bank can reliably replace the cash it no longer keeps on hand.
With a hypothetical $1 billion buffer, that’s $80 million to $170 million potentially available for other uses, assuming regulators and risk managers allow it. That amount isn’t profit, and keeping less cash on hand means the bank must be more confident it can borrow on short notice.
The day your lender wants more bonds
Even perfectly coordinated records become less impressive when the lender changes the terms. Repo lenders usually advance a little less cash than the securities are worth, keeping a cushion in case the borrower defaults or the bonds fall in value.
Traders call that difference a haircut, and its size determines how much money the borrower can actually obtain against their portfolio.
Put $1 billion in government bonds into a hypothetical repo agreement with a 2% haircut, and the lender advances $980 million. Increase the haircut to 5% while the bonds’ market value stays the same, and the advance falls to $950 million.
The dealer now has $30 million less cash than before, despite owning the same securities and using the same settlement technology.
Original illustration, using actual/360 interest accrual, unchanged bond prices, and no extra fees. The lower cash advance and the higher interest expense are separate costs.
Keeping the old $980 million loan under the new 5% haircut would require about $1.0316 billion in eligible bonds, or $31.6 million more in collateral. Broadridge could help move those extra securities faster if the dealer owned them and the lender would accept them.
Yet, the software can’t produce the missing bonds, and it can’t force a money market fund or another bank to part with cash simply because the transfer will settle quickly.
Things also get difficult if the lender refuses to renew the deal. Repo is short-term financing that must be arranged repeatedly, making borrowing for a couple of hours efficient when lenders are plentiful and very difficult when they all want their money back.
The dealer has to find another lender, draw on available funding, or sell assets, possibly at a price it wouldn’t have accepted the week before. Faster transfers help execute whichever option exists, but they don’t supply the option.
The Federal Reserve has studied how short maturities and the reuse of collateral can connect borrowers and lenders through chains of obligations. Securities received in one transaction can secure another, so a missed repayment can inconvenience more than the two original participants.
Wall Street is rebuilding the plumbing, not abolishing the plumber
Broadridge’s new technology is emerging just as regulators try to overhaul another part of the same market.
The SEC has set Dec. 31, 2026 as the compliance date for eligible cash Treasury transactions under expanded central-clearing requirements, with covered repo transactions due by June 30, 2027.
Those deadlines have generated plenty of industry work, but central clearing is not another name for blockchain settlement.
Settlement is when the agreed money and securities are actually delivered. Central clearing places an institution, known as a central counterparty, between the parties to eligible trades, allowing it to offset obligations, collect margin, and manage certain defaults under its rulebook.
The two functions can work together, but software that moves a token and confirms a payment doesn’t automatically become the institution standing behind the contract.
An analysis of settlement on Solana addressed a similar distinction: delivering two assets together doesn’t create the credit needed to finance the transaction in the first place.
The Office of Financial Research found that about 46% of Treasury repo volume in its early-2026 observations was centrally cleared, compared with roughly 62% of the narrower category subject to the mandate.
Among the largest banks, the cleared share rose to about 55% at the end of 2025 before returning toward 45%, while other banks moved in the opposite direction. Those figures track which trades go through a central counterparty, not which banks use Broadridge.
New OFR research published on Oct. 20 offers a particularly good illustration of why activity and economic effect are different measures.
Under a backward-looking simulation that assumes traders don’t alter their behavior, the expanded clearing mandate would add slightly more than $500 billion in daily transactions to the volume used to calculate the SOFR overnight rate while moving the rate itself by only about one basis point, or 0.01 percentage points.
It shows that moving a huge amount of financial activity into a different process doesn’t necessarily change the cost of borrowing.
Bitcoin transfers don’t require a repo desk, but traders financing risky positions can feel the effects when lenders conserve cash.
Stablecoin issuers including Circle and Tether hold Treasuries among their reserves, while Circle’s USDC depends on assets and payment arrangements outside crypto.
None of those connections means Broadridge’s transaction volume moves Bitcoin’s price, or that every blockchain repo makes banks more fragile. However, they do help explain why a seemingly obscure market for borrowing against bonds deserves attention outside the trading desks that use it.
When a bank needs $1 billion at 9 a.m., the new system can deliver collateral and repay the loan by 11, wasting less time, money, and paperwork.
If the expected payment doesn’t come in on time and lenders stop answering, the bank still has to find the dollars somewhere, however quickly its blockchain can update the books.
