instant

$4 trillion of US stocks change hands every day, on rails built for a nine-to-five market.

$4.0 tn
of US stocks traded every day
386 m trade sides · NSCC, Q1 2026
3.77 m
holders of tokenised stocks, +78 % in 30 days
$3.15 bn held on-chain · rwa.xyz, 22 Sep 2026
53 h
a week with no US stock market, even after Nasdaq goes 23/5
Fri 20:00 to Sun 21:00 ET, plus four daily pauses · from 6 Dec 2026
T+1
from trade to ownership in the US. On-chain: seconds
and the share is collateral the moment it lands

Why the market wants on-chain: four reasons

  • Open when clients are. 24/7/365. Incumbents stretch their clocks (NSCC clears 24x5 since 29 Jun 2026, Nasdaq trades 23x5 from 6 Dec 2026) but stop at the weekend, and Fedwire won't move dollars on a Sunday before 2028–29.
  • Collateral at once. A share settled on-chain can be pledged the same second, anywhere, with no custody chain. DTCC built its own pilot for "real-time collateral mobility".
  • One ledger, less reporting. Buyer, seller, custodian and regulator read the same record, so reconciliation and post-trade reporting shrink to a lookup.
  • Global by default. Clients fund in stablecoins from anywhere, with no US broker-dealer chain in between.

Why now: the rails and the demand are moving

  • DTCC ran real tokenised stock trades on 15 Jul 2026 with 30+ firms; public launch Oct 2026. On 17 Sep 2026, the SEC let tokenised US stocks trade in on-chain liquidity pools for five years.
  • Demand is retail and global: Robinhood has sold 24/7 stock tokens in 120+ countries since 1 Jul 2026; Nasdaq got SEC clearance to trade tokenised shares on 18 Mar 2026.
  • On-chain dollars at scale: $306 bn of stablecoins, $7.5 tn of tokenised repo settled in June 2026.

The opportunity is not cheaper settlement: it is the clients, hours and assets this system cannot serve.

Sources: NSCC CPMI-IOSCO disclosure Q1 2026; SEC approval of Nasdaq extended hours (2026) and Nasdaq launch notice; DTCC, NSCC 24x5 clearing (29 Jun 2026); Federal Reserve, Fedwire operating days; SEC Release 34-105047 (18 Mar 2026); DTCC 15 Jul 2026; SEC innovation exemption (17 Sep 2026); Robinhood (1 Jul 2026); rwa.xyz (22 Sep 2026); DefiLlama 8 Sep 2026; Broadridge 7 Jul 2026.

scroll

No clearing house means no credit: every trade must be paid in full, instantly. We finance that instant.

The catch: on-chain, settlement is atomic

  • No one stands in the middle. A public chain has no clearing house: no netting of buys against sells, no one vouching for a stranger's ability to pay. Delivery and payment happen in the same second, or not at all.
  • So every purchase is paid 100 %, at execution, in dollars already on-chain. The client's cash still travels on bank rails, and lands a business day later.
  • The broker carries the gap. Either it parks a stablecoin float sized to its busiest day on its own balance sheet, or it turns the order away.

Our protocol: one smart contract, one atomic transaction

  1. 1

    A flash loan advances 100 % of the purchase amount

  2. 2

    Atomic delivery-versus-payment: cash out, tokenised share in

  3. 3

    The share just bought is pledged as collateral in the same transaction

  4. 4

    The collateralised loan repays the flash loan

All four steps commit together or none of them happened. Then, outside the transaction: the secured loan runs for a few hours, up to a day, until the broker repays it from its client's cash.

100 %
of every purchase paid at execution, on-chain
no netting, no credit, no next day
~70×
more cash to fund the same trades than through the NSCC
the NSCC nets 98.6 % of what it clears; on-chain, every dollar moves · market-wide average

How it works  ·  where the liquidity comes from

Capital from anywhere, risk curated by professionals.

  1. LendersIdle capital, from anywhere in the world.
  2. Curated poolsProfessional curators set the limits it lends within.
  3. Our protocolRoutes the flash loan: 100 % of the price, from the pool to the broker.
  4. The brokerSettles atomically, repays with a fee, and the yield returns.

Question I  ·  Is the data corrupted?

The data isn't corrupted, it's just not settled yet. People pay for that one-day gap.

Who pays for the gap

  • Retail investors: trading gets frozen when the system needs cash fast, not when they have done anything wrong.
  • Small and non-US brokers: shut out entirely, not because their trades are risky, but because the entry ticket is too high.
  • Everyone else: billions immobilised at all times, just in case, money that cannot be used while it waits.

And blockchain cannot inherit the fix

  • As we said: no company stands in the middle of a public chain, so no one can vouch for a stranger's ability to pay.
  • The safety net disappears with the middleman: the buyer must pay in full, at execution.
  • That is the constraint our protocol exists to answer.

The cost of the gap, in numbers

$16.2 bn
immobilised at the clearing house at all times
31 Mar 2026 · and $33.5 bn on the morning of 28 Jan 2021
95 %
of that collateral is held in cash
the most expensive form there is; equities are not accepted
3 of 149
foreign firms among NSCC participants
31 Mar 2026 · plus a $250,000 minimum deposit before the first trade

The record isn't wrong. It is simply unfinished for a full day, and someone has to be paid to carry that day.

Question II  ·  Do outsiders have an incentive to corrupt it?

We advance a stranger the full purchase price. The incentive to disappear is real, so we remove the opportunity, not the motive.

Four moments, four temptations

  1. 1

    We send the full amount. Keep it, never buy.

  2. 2

    They buy the stock. Keep the stock, skip repayment.

  3. 3

    The stock becomes collateral. Pledge it twice, to two lenders.

  4. 4

    Collateral repays the loan. Hand over something worthless instead.

One move, not four

  • The old system prevents this by vetting who is allowed to trade at all.
  • We prevent it differently: the four steps are one transaction. They commit together, or none of them happened.
  • There is no midpoint at which anyone can walk away holding only part of it.

The atomic construction removes the risk of the settlement instant. What survives the transaction is an ordinary secured loan.

For the seconds of the advance we lend to a stranger with nothing held against it, and it is still safe. Not because we trust them, but because the alternative cannot execute.

Trust does not vanish. Someone still sets the lending limits, something still has to price the collateral, and the hours that follow the trade are a secured loan. That residual window is exactly what curators are paid to price. We removed trust from the one place that used to cost billions, not from everywhere.

Question III  ·  How much does the business depend on others trusting the data it is built on?

Entirely. NSCC asks you to trust one company. We ask you to trust four data feeds. That is the whole trade, and it cuts both ways.

What trust used to look like

  • NSCC produces trust the old way: vetting members, holding $10 to 20 bn of collateral, 50 years of track record. Lehman, 2008: $500 bn unwound, zero loss to the fund.
  • You never verify NSCC's data. You trust the institution, and the SEC supervises it for you.
  • We remove the institution. Every ounce of trust it produced now has to come from data that anyone can check.

The four feeds we live or die on

  • The ledger

    The token really moved and the transfer is final.

    Breaks ifa court or insolvency administrator refuses to recognise on-chain finality.

    Safeguardsettlement-finality law (EU proposal, Dec 2025), permissioned or designated chains.

  • The token

    The tokenised share is a real share, held by a registered transfer agent.

    Breaks ifit's a synthetic wrapper with no rights (STA flagged ~$2 bn of such tokens to the SEC, July 2026). Then our collateral is worthless: Danielle's step 4.

    Safeguardaccept only onshore, transfer-agent-backed tokens.

  • The price

    The oracle valuing our collateral is honest for every second of the loan.

    Breaks ifthe feed is manipulated. Flash loans are the number one tool for oracle attacks: Euler $197 m (2023), Balancer $116 m (Nov 2025, after 10+ audits); about 20 % of all value stolen in DeFi.

    Safeguardmulti-source oracles (DTCC chose Chainlink for its own collateral chain, May 2026), conservative haircuts, time-weighted prices.

  • The dollar

    The stablecoin we lend is worth $1.

    Breaks ifreserves fail. USDC traded at $0.87 for a weekend in March 2023.

    Safeguardregulated issuers, tokenised bank deposits (JPMD), central-bank-backed cash (Fnality).

Trust dependency map

FeedWho produces itWhat breaksHow verifiableResidual trust needed
LedgerThe chain's validatorsLegal finality of the transfer Public
TokenIssuer and registered transfer agentCollateral with no rights behind it Attestation
PriceOracle networkCollateral mispriced, pool drained Public
DollarStablecoin issuer or bankThe loan is worth less than $1 Audit

The paradox, in numbers

98.6 %
of flows NSCC nets by being trusted
the trust we have to replace
~20 %
of DeFi losses come from flash-loan-enabled attacks
the tool we use is the tool used against data
$500 bn
unwound by DTCC in the Lehman failure with zero loss
the benchmark our data has to match

Trust doesn't disappear on a blockchain. It moves from one institution you cannot audit to four data feeds you can. Our business is not a lending business; it is a data-verification business that happens to lend.

One feed being wrong for twelve seconds is enough to drain a pool. The institution fails slowly and visibly; data fails fast and silently. That asymmetry is the price of removing the middleman.

Conclusion  ·  three questions, one honest answer

The data isn't corrupted, the incentive to corrupt it is real, and we depend on it completely. So the product only works where the clearing house doesn't.

What the three questions told us

  • Q1: The record is right, it's just a day late. That day costs $10 to 20 bn in locked collateral and, in a crisis, a broker's ability to trade.
  • Q2: A stranger has every reason to run with the money. Atomicity removes the opportunity, not the motive: all steps complete together or none do.
  • Q3: The business is 100 % dependent on four data feeds being trustworthy. It replaces institutional trust with verifiable data, and inherits data's failure modes.

Where it works, and where it doesn't

  • Doesn't work as a replacement for NSCC. NSCC nets 98.6 % of flows; we would fund 100 % of them. Interest on 50 to 70 times more principal beats any saving on fees. "Faster is not always better" (SIFMA, June 2026) is arithmetic, not lobbying.
  • Works where NSCC is absent. Tokenised stocks ($10 bn+ traded via Backed, $1 bn+ on Ondo), 24/7 and weekend trading, non-US and small brokers who can't meet the $250 k plus capital plus legal bar. There, the alternative isn't netting; it's no market at all.
  • Works better with a clock. Most of netting's benefit is captured in 10-minute to 1-hour windows (Journal of Financial Market Infrastructures). Batch trades inside a window, fund the gross with the flash loan, net inside the batch, settle atomically. That is netting without a clearing house, and it's our next step.

What must be funded per $100 traded

NSCC bar approximate: derived from the 98.6 % netting rate and the VaR charge of $8.4 bn on roughly $3 to 4 tn daily notional. The batch figure depends on the flow inside the window and has yet to be measured.

The market we're actually in

$4 tn / day
cleared by NSCC
not our market
~$26 bn
all tokenised real-world assets today (excl. stablecoins)
our market, small and growing
11 Oct 2027
Europe moves to T+1
the day the funding gap becomes a European problem too

We don't replace the clearing house. We finance the trades it was never built to clear.