Atomic settlement swaps one risk for another, and banks aren’t ready

Batch settlement has long been the banking industry’s pact with time. It let markets trade at full speed while finality waited for cutoffs, netting cycles, correspondent windows and, often, the next business day. That delay did not remove exposure; it stored it.

Processing Content

Tokenization breaks that pact. When settlement becomes atomic, time stops cushioning the system, and the ability to fund, collateralize and synchronize becomes the real gatekeeper of execution. I have watched this shift reshape the demands on bank treasury operations in ways most institutions aren’t yet prepared to handle.

“Atomic” has a precise meaning: Either every leg of a trade settles, or none does. Bank for International Settlements, or BIS, publications, drawing on the CPMI glossary, describe atomic settlement as the linking of two assets, so that a transfer occurs only if the other one does too. In delivery-versus-payment and payment-versus-payment settlement structures, this eliminates principal risk because neither side can be left holding an unmatched obligation. 

U.S. regulators have reached similar conclusions. The FDIC specifically cites reduced loss exposure during the settlement window, particularly in cross-border or counterparty-unknown scenarios.

But here’s the trade-off that matters most for banks: Instant settlement can raise liquidity needs relative to netted arrangements, and it makes liquidity management more complex. Atomicity replaces uncertainty about completion with uncertainty about funding sufficiency.

Batch settlement allowed settlement to look like an end-of-day accounting problem. Atomic settlement forces it to behave like a real-time balance-sheet constraint. Credit exposure created by lags can shrink, but liquidity risk becomes immediate, binary and inseparable from execution.

This rewrites accountability inside banks. Under batch, post-trade teams had room to reconcile, borrow, pledge or net away problems before finality. Under atomic settlement, treasury policy, collateral mobility and intraday funding capacity start deciding what the front office is permitted to execute.

If risk now lives on the balance sheet, visibility becomes nonnegotiable.

Regulators have spent years dragging intraday liquidity out of the basement. The Basel Committee on Banking Supervision’s monitoring tools were designed to help supervisors gauge intraday liquidity risk and a bank’s ability to meet payment and settlement obligations on time, under both normal and stressed conditions. European supervisors are explicit that “good” means real-time visibility: monitoring payment flows as they happen, with alerts for unexpected needs, and seeing balances, projections, collateral, and credit lines across relevant systems.

Tokenization turns that supervisory ideal into an execution requirement. Consider what’s already happening: Clients now park cash buffers across regions, then use tokenized transfers to move funds 24/7/365, including during market closures and holidays. Citi’s recent positioning of its tokenized payment infrastructure as enabling round-the-clock, multibank cross-border instant payments signals where institutional liquidity management is heading.

In an always-on environment, “intraday” stops being a subcategory. It becomes the operating condition.

The quiet power of batch settlement was netting. It reduced gross activity into smaller cash movements and made liquidity look cheaper than it was. Atomic settlement does not inherit that benefit automatically.

If a market wants liquidity efficiency, it has to recreate netting explicitly through orchestration layers, liquidity-saving mechanisms or pricing models that treat netting as a service instead of an assumption. Otherwise, the system gets the BIS’ downside, fewer settlement fails, but higher liquidity needs and more complex liquidity management.

Central banks aren’t waiting for the industry to figure this out.

The BIS and the Bank of England have already tested what atomic means in production-minded designs. Project Meridian demonstrated “synchronization,” where funds move in real-time gross settlement, or RTGS, only if a corresponding asset on another ledger moves at the same time, orchestrated by a synchronization operator. Subsequent phases tested synchronization across FX settlement between RTGS systems and explored atomic settlement for tokenized securities with programmable triggers for liquidity management.

The implication is blunt: The historical separation between trading, treasury and settlement is being engineered out of existence. Atomic settlement only works when funding and collateral constraints are enforced at the moment of execution.

But speed without resilience is a liability.

Speed amplifies whatever the system already is. The FDIC has warned that instant settlement has the potential to increase the speed and intensity of runs. In a failure scenario, regulators may require an “off switch” to stop tokens moving immediately at the point of a bank failure.

That warning should land with any bank thinking about always-on settlement. Operational resilience becomes settlement resilience, because an outage or a data error is not a delay anymore, it’s a finality problem.

It is also why the public narrative around lightning-fast atomic settlement can be misleading. Crypto coverage increasingly treats atomicity as a synonym for speed and inevitability, mapping central-bank experiments onto public ledgers.

Banks don’t get to treat atomic settlement as marketing. They have to treat it as a new risk geometry.

Tokenization does not eliminate settlement risk. It moves it from delayed counterparty uncertainty into immediacy, funding precision and system integrity. When time is removed as the buffer, liquidity becomes the shock absorber. And it must be engineered, priced and governed accordingly.

Banks that approach atomic settlement as “batch, but faster” will find it unforgiving. Structural advantage will accrue to institutions that can see positions as they form, mobilize collateral in real time and embed funding constraints into execution instead of reconciliation.

The time for stress-testing intraday liquidity frameworks against atomic settlement assumptions is now, not after the first live failure.

Atomic settlement is not a technology upgrade. It is a redesign of how the balance sheet participates in markets. The banks that accept that early will define the next settlement standard.