I can send my brother a photo from a moving train and it lands before I have put my phone back in my pocket. It costs nothing. It does not care which phone he has, which country he is in, or what the time is. If I owe him twenty dollars, the transfer is free now too, and it will usually land the same day, which retires most of the old complaint about fees and overnight waits. What is left is narrower and stranger. I cannot hand him the money in the moment, the way the photo arrived in the moment. It does not go to him, wherever he is, knowing only who he is; it goes to an account number, when the system is ready. The photo and the twenty dollars travel the same distance, between the same two people, and only one of them behaves like it is 2026.
This is not abstract for me. I was the first hire at Dolla, a New Zealand company built to make money move like a message: real-time payments between people, and cheaper payments to merchants, running over open banking through Akahu instead of the card networks. So I am not speculating about whether it can be done. I helped build the team that did it. What I want to write about is the part I did not expect.
Money is harder than messages
The obvious reading is that money is simply harder than messages. It has to be secure. It has to be final. There is fraud to stop, regulation to satisfy, and two institutions that have to agree before anything settles. Of course it is slower. It will catch up once the banks finish upgrading their systems.
Every clause of that is true, and it is still not enough, because the rail it says we are waiting for already existed. In New Zealand you could connect a bank account to an app through a provider like Akahu and move money directly, bank to bank, with no card network in the middle. That is the rail Dolla ran on. The pipe was built. The standards were written. What I did not think hard enough about at the time is that nothing required any bank to stay connected to it. Access in those years was voluntary, negotiated arrangement by arrangement, and a voluntary door can be narrowed, or queued, or closed at the pleasure of whoever owns the doorframe. A direct transfer that costs almost nothing was never a technology anyone was waiting on. It was a technology waiting to be allowed at scale.
Who laid the road
I went in assuming the hard part was the product, or the rail underneath it. Build the better experience on the better pipe, I thought, and people will come.
Here is what I had wrong, and unlike most lessons from inside a startup, this one is checkable from the outside. Everywhere money has actually started to move like a message, the sequence is the same, and it never starts with a bank volunteering. In 2016 the United Kingdom’s competition regulator ordered the nine largest banks to build shared open-banking standards, live from 2018. (Open Banking Tracker 2026) The European Union’s PSD2 rules took effect in January 2018 and required every bank in the bloc to open access to licensed third parties. (Open Banking Tracker 2026) Australia was given open banking by government scheme, the Consumer Data Right, not by a bank’s offer. (Cuscal 2026) In all three, a regulator laid the road. New Zealand spent those same years running the voluntary version: access that existed only where a bank had agreed to it.
Then New Zealand proved the rule from the other direction. The Customer and Product Data Act passed in March 2025, and banking was the first sector designated under it: from December 2025 the four biggest banks were required to open standardised access, with Kiwibank phased in from June 2026. (Ministry of Business 2026) (Russell McVeagh 2025) Akahu, the provider Dolla was built on, was accredited as an intermediary under the new regime on 19 December 2025. (Akahu 2025) Every date in that paragraph sits after Dolla had already wound down. For the whole of the company’s life, the road it needed was private, provisional, and lent. The mandate that would have made it public arrived once there was no longer a Dolla to drive on it.
That ordering is not bad luck, or not only bad luck. Ask a bank to help make instant, near-free, bank-to-bank payments the norm and you are asking it to give up the fees, the float, and the card relationships it earns while money moves slowly through channels it owns. A toll booth does not campaign for a new road that bypasses the toll. It does not even have to refuse one. It only has to be unhurried.
The toll booth between two people
Between you and a friend who owes you twenty dollars sits a middle layer: a card network, a bank, or a payment app that is one of those two wearing a friendlier face. Every option marketed as instant is instant because someone in the middle fronts the money and takes something for standing there. Between two friends the toll is mostly taken in time and clumsiness. Where it is taken in money is at the counter: about $1 billion a year is what card fees cost New Zealand businesses, on the Commerce Commission’s own figures, and the Restaurant Association describes the fees hospitality pays as among the highest in the OECD. (RNZ 2025) Those numbers are not laws of nature, and the proof is what happens when a rule touches them. When the regulator capped interchange fees under the Retail Payment System Act, the first caps alone saved businesses about $140 million a year. (Commerce Commission 2026) A price that falls the moment a rule changes was set by somebody, and it can be set again.
The delay side of the toll is harder to pin down, and I want to be honest about the difference. While a payment clears, the money sits somewhere, and whoever it sits with is earning on it, so every hour in the middle is an hour someone in the middle is paid for. But nobody publishes a line item isolating what slowness earns, so I cannot give you that number, and I would distrust anyone who could. Take it as an argument assembled from how interchange and float work, not as a fact: money moves exactly as fast as it is profitable for the people in the middle to let it move, and no faster.
Until the challenger runs out of time
None of this looks like a fight from the outside, which is why the story is easy to misread. The most powerful incumbents rarely kill a better idea by refusing it. They kill it by being slow. They study it, they pilot it, they raise reasonable concerns, and the clock that costs them nothing keeps ticking while a small team burns the runway it does not have. Delay is the cheapest weapon an incumbent holds, and against a challenger it is often enough. A mandate is the one thing that takes the clock out of the incumbents’ hands.
What I will not claim
Dolla is not trading today. I left before the end, and I am not going to reduce a company’s wind-down to a single cause I was not there for, because there is never one reason and it would be too convenient for the argument I am making. Blaming the incumbents would round the story off neatly, and I do not know that it is true. The mandate’s timing is a fact about New Zealand, not a verdict on one company.
What I do know, from the inside rather than from a diagram, is that the pace problem is real. The rail can exist, the product can be good, the team can be right, and the thing can still be slowed, by the parties who profit from slow, for longer than a small company can wait. When I say money could move like a message, I am not describing something impossible. I have seen it move. Whether it moves for everyone was never a question about code, or even about any one company. It is a question about who is allowed to set the speed, and New Zealand has only just taken that decision away from the people who profit from slow. The road is public now. Nobody who owned a toll booth built it.