top of page

Where the Toll Booth Goes Next

15 hours ago
5 min read
Where the Toll Booth Goes Next

By Arunjay Katakam, Author


For years, I argued that open financial infrastructure would make moving money cheaper. Remove the unnecessary intermediaries, settle faster, release the capital trapped in correspondent accounts, and the savings would reach the people making the payments.

I still believe much of that. What I no longer believe is that removing an intermediary necessarily removes its margin.


Money is becoming software, and the question is what that does to everyone who sits in the middle of it. Champions of open finance say it ends the middleman. Critics fear the same thing. However, look at what actually happened in 2025. The GENIUS Act passed, Stripe bought Bridge and Circle listed on the New York Stock Exchange. The middleman hasn’t disappeared. Traditional finance and crypto sat down to negotiate the terms of a convergence, and the question was who would collect the toll next.


The toll matters. The World Bank puts the global average cost of sending a remittance at 6.36 percent, more than twice the UN target of three percent. [1] A family sending money home pays for far more than the movement of funds. It pays for access, liquidity, compliance, distribution and the margins accumulated along the way.


I used to describe that mainly as an efficiency problem. It is also a question of market power. Building a faster rail may reduce the cost of a payment, but who controls access to that rail and who captures the savings are still open questions.


The toll booth moves


Consider liquidity. A shared settlement pool may let licensed participants avoid maintaining separate accounts with correspondent banks. That is useful, but an illiquid currency pair remains illiquid. Someone must provide the funds on the other side of the trade, bear the risk and earn a return for doing it.


The cost can become easier to see and compete over, but it does not disappear. Even atomic settlement requires assets to be available when the exchange takes place; its liquidity needs differ from those of correspondent banking, but they are still real.


The internet offers a familiar warning. It made access to information radically more open, then gave rise to companies with enormous power over discovery and distribution. Finance could follow a similar path. The valuable position may move from the correspondent account to the wallet, the stablecoin issuer, the identity provider or the AI agent that chooses which offers a customer sees.


That is a different proposition from the one I spent a decade making. Open infrastructure creates the possibility of competition. It does not guarantee the result.


What the bank is left holding


Banks have reason to worry about where the customer relationship goes. If a customer holds a tokenised bank deposit in a wallet operated by another company, the bank may retain the liability while the wallet provider controls the experience. If that customer exchanges the deposit for a stablecoin, the bank may lose the funding as well.


Swift’s response is a shared ledger announced in July 2026, with 17 banks preparing to pioneer cross-border payments using tokenised deposits. The ledger coordinates transfers between bank-issued deposits; final settlement still takes place through existing systems. Banks are moving to make their money usable in a more open environment while keeping a role in the system built around it. [2]


Another response is Open USD. Its proposed model brings together more than 140 businesses and shares reserve earnings with participating partners. That may prove attractive to firms that have had little share in stablecoin economics. It also raises a question that should be asked of every network that achieves scale: who sets the rules for joining, earning and leaving? [3]


The names attached to the initiative include some of the largest companies in payments and technology. That does not make it a closed system, however it does mean that openness should be judged by its terms of access and governance, rather than its name or the number of partners at launch.


Custody is a service you earn


Banks can compete in this world. But they will need a better answer than making departure difficult.


Revolut offers one example. It introduced crypto trading in its app in 2017 and now supports withdrawals to external wallets for eligible assets. Letting a customer take assets elsewhere gives up a measure of control. It also forces the provider to give people a reason to stay. [4]


That principle reaches beyond crypto. If people can move their money and credentials without losing access to services, providers have to compete on the quality of those services: advice, reliability, price and the way they treat customers when something goes wrong.


Bankers rightly point out that payments infrastructure also pays for security, compliance and resilience. I agree. The aim should not be to strip every safeguard from a transaction. It should be to ask which charges pay for a valuable service, and which persist because the customer has no practical alternative.


Governing the new chokepoints


If the toll moves rather than disappears, the focus shifts to governance. A wallet can make switching hard. An identity provider can determine who is recognised. A stablecoin issuer can set the terms of redemption. An AI agent can decide which products are considered before its owner sees a choice at all.


I would start with three expectations for providers that become essential gateways: clear and fair access terms, usable ways for customers to leave with their money and credentials, and transparent pricing. Regulators should keep looking across the whole transaction to see where value is captured. Otherwise, a fall in payment fees at one layer may simply reappear as a charge at another.


AI agents may unsettle an old source of bank power. An agent comparing providers for its owner has no childhood memory of a branch and no affection for a logo. But agents will have defaults, rankings and commercial relationships of their own. We should pay attention to who sets those, too.


The moral case for open finance remains straightforward. The people least able to afford the toll often pay the most. I once thought winning meant removing the toll booth. Now I think it means making every toll visible, open to challenge and answerable to the people who must pass through it.

When money becomes software, the central question is who controls the financial operating system.


Arunjay Katakam is the author of The Future of Finance is Open: A Visionary Tale of the Great Convergence, which launches at Money20/20 USA in Las Vegas this October and at The Singapore Fintech Festival this November.


[1] World Bank Remittance Prices Worldwide Issue 54, Q3 2025 https://remittanceprices.worldbank.org/

[2] Swift Shared Ledger

 
 
bottom of page