What Happens When the US Settlement Layer Goes On-Chain

The Clearing House settles more than $2 trillion every business day. That's CHIPS, the interbank wire system, plus EPN for ACH, plus the RTP real-time payments network. Twenty-five of the largest US banks own it. When this institution decides to put money on a blockchain, the question changes from "will banks adopt tokenized deposits?" to "what happens to the settlement architecture that already works?"

On September 24, TCH announced it had selected Quant, a UK-based programmable money infrastructure company, to power its On-Chain Money Initiative. The network will let US banks clear and settle tokenized deposit transactions, with connectivity to existing fiat rails including CHIPS and RTP. It's expected to launch in the first half of 2027.

But this isn't a bank experimenting with a token on a test chain. This is the operator of the actual US interbank settlement layer adding a new clearing mechanism alongside the ones that already move real money. And the architectural choice TCH made, to settle the token after the fiat rather than replacing the fiat leg entirely, tells you more about what on-chain settlement will actually look like in the US than any whitepaper does.

What CHIPS does that blockchain doesn't

CHIPS handles wire transfers between banks. A corporate treasurer at JPMorgan sends $40 million to a supplier's account at Citi. That payment enters CHIPS as a gross obligation, one bank owes another $40 million.

But the payment doesn't settle as a $40 million transfer. CHIPS runs multilateral netting across all of the day's obligations. If JPMorgan owes Citi $40 million and Citi owes JPMorgan $35 million, the net obligation is $5 million in one direction. Multiply that across 25 banks and thousands of daily payments, and the netting ratio typically exceeds 95%.

Through August 2026, CHIPS processed $377 trillion in payment value, averaging $2.245 trillion per business day. That's the gross number. The net settlement, what banks actually transfer to close the day, is a fraction of that. Netting is why the system works. Without it, banks would need to prefund $2 trillion in daily liquidity instead of the much smaller net amount.

On-chain settlement works differently. A tokenized deposit moving on a blockchain settles atomically, the sender's balance decreases and the receiver's balance increases in the same transaction. There's no netting window. Each transfer is final the moment it confirms.

That's a feature in some contexts. Delivery versus payment in securities transactions benefits from atomic settlement because the asset and the cash leg need to move together. But for bulk interbank payments, where the whole point is to reduce gross obligations through netting before settling the difference, atomic settlement would be a step backward. Banks would need more liquidity, which defeats the purpose.

How TCH chose to connect the two

TCH's design resolves this by keeping the two finality models separate. The tokenized deposit clears on-chain, but it doesn't settle there. According to Quant's October perspective piece, "the tokenised deposit is delivered once the fiat has settled over RTP or CHIPS."

So the sequence runs like this: the bank initiates a tokenized deposit payment. The clearing happens on the shared permissioned ledger that Quant is building. But the actual money movement, the interbank settlement, runs through the same CHIPS or RTP infrastructure it always has. Only after the fiat leg confirms does the on-chain token move to the recipient.

This means netting survives. If ten tokenized deposit payments flow between JPMorgan and Citi in a day, CHIPS can still net them against each other and against the regular wire payments in the same clearing window. The on-chain layer adds programmability and richer data to the payment, but it doesn't touch the settlement mechanics underneath.

It also means the on-chain token functions as a receipt. The token confirms that the payment happened and what conditions triggered it, but the money moved on fiat rails. The blockchain carries the instruction and the proof, while CHIPS and RTP carry the cash.

What the UK did first

Three weeks before TCH picked Quant, the UK ran the world's first interbank tokenized deposit transactions. Seven banks, Lloyds, NatWest, Barclays, HSBC, Monzo, Nationwide, and Santander, completed live remortgage and marketplace payment transactions using tokenized deposits on Quant's Overledger platform.

The settlement model was different from TCH's design. The UK's Great British Tokenised Deposit initiative settled over the Bank of England's RTGS, the country's central bank real-time gross settlement system. That's a different architectural choice: settling through the central bank rather than through an interbank netting system.

The UK banks chose an omnibus control account that always matches the total value tokenized. Quant's perspective piece notes this "simplifies settlement and scales well," while the alternative, a shadow account per customer, "becomes hard to manage beyond a few thousand corporate clients."

What matters for the US comparison is the use cases both programs are targeting. The UK ran remortgage completions where funds release automatically once the property transfer is confirmed, and marketplace payments where the buyer's money is held until delivery is verified. Both are conditional payments, transactions that trigger when a condition is met.

Those conditions are the point of putting deposits on a blockchain. A wire through CHIPS carries an amount, a sender, and a receiver. A tokenized deposit can carry the conditions under which the payment executes: release the funds when the title transfer records, or pay the supplier when the shipment arrives at the warehouse. The programmability rides on the token while settlement stays on the existing rails underneath.

What Quant builds

Quant's role is the orchestration layer. Each bank runs its own stack for issuance, wallets, and custody, connected to a shared permissioned ledger that Quant coordinates. Quant handles the interoperability between banks and the connectivity between the on-chain layer and the fiat rails.

Quant also offers Tokenized Deposits-as-a-Service for banks that don't have their own on-chain infrastructure. TCH's announcement said the initiative would be "accessible to financial institutions across the United States, enabling banks of all sizes to participate." The TDaaS model means a community bank doesn't need to build a blockchain stack to issue tokenized deposits on the network.

This matters because most tokenized deposit experiments so far have been single-bank systems. JPMorgan's JPM Coin moved money between JPMorgan accounts, and Citi's Token Services stayed inside Citi. The deposits couldn't cross institutional boundaries because each bank built its own walled garden. TCH's network is the interbank layer that connects them, which is exactly the role TCH plays for wires and ACH today.

The June announcement listed quotes from Bank of America, Citi, JPMorgan, Wells Fargo, HSBC, and a dozen other banks. The September selection of Quant confirmed the technology partner. The network is expected to go live in the first half of 2027.

The settlement models compared

Three major tokenized deposit initiatives are now running or building, and each chose a different settlement model.

The UK's GBTD settles over the Bank of England's RTGS using central bank money as the settlement asset, while Swift's shared ledger handles clearing and leaves settlement to existing correspondent banking arrangements. TCH settles over its own rails, CHIPS for wires and RTP for instant payments, keeping multilateral netting intact.

What none of them chose is fully on-chain settlement where the token itself is final. That model exists in DeFi and in some stablecoin transfers, but every bank-led initiative keeps the final money movement on a fiat rail, using the blockchain for instructions and conditions while the existing infrastructure moves the cash.

BIS's Project Agora takes a step further by putting central bank money on jurisdiction-specific ledgers so that settlement can be atomic, completing on-chain in the same transaction. But Agora depends on which central banks participate, and neither the Federal Reserve nor the Bank of England has committed to putting central bank digital currency on a shared ledger for interbank settlement.

What this means for the US

TCH's On-Chain Money Initiative is designed for corporate treasury, liquidity management, cross-border payments, digital asset settlement, and agentic commerce applications. An AI agent executing a conditional payment needs the conditions to live somewhere the recipient can verify, and a programmable token carries those conditions in ways a CHIPS wire message can't.

The GENIUS Act, which takes effect January 18, 2027, creates a federal framework for stablecoin issuers but treats tokenized deposits as ordinary deposits under existing banking law. That regulatory asymmetry gives tokenized deposits an advantage: they don't need a new license, new capital rules, or new reserve requirements. They're deposits, and the banks already have the charter.

But the two-finality-model design means tokenized deposits won't replace CHIPS or RTP. They'll ride on top of them, keeping the settlement layer in place while turning the instruction layer from a flat message carrying an amount and a counterparty into a programmable token carrying conditions, data, and automated workflows.

The UK proved the concept works between banks in September. TCH is building the US version for 25 of the country's largest institutions, with a technology partner that already ran the UK pilot. The architecture tells you where this is going: the blockchain becomes the instruction layer, while the existing infrastructure stays the settlement layer.