Sovereignty Is Built at the Moment of Use

On a beach in Brazil, I bought something from a vendor and paid him by scanning a QR code on his phone. The transfer cleared in about ten seconds. He didn’t need a card terminal. He didn’t pay a percentage to anyone. And at the end of the day he wouldn’t be walking home with a pocket of cash, which in a lot of places is the single most dangerous part of running a small business.

The system is called Pix. Brazil’s central bank built it and launched it in 2020. It now has more than 170 million users, roughly 93% of the country’s adult population. Around 60 million Brazilians have no credit card at all, and Pix works for every one of them.

I have been thinking about that experience ever since, because it cuts against how we usually talk about sovereignty.

The sovereignty conversation, in Canada and almost everywhere else, is a conversation about ownership. Who owns the infrastructure. Where the data physically sits. Whose rails our payments, our clouds, and our critical systems ultimately run on. These are real questions and I don’t want to wave them away. But they are questions about procurement, and procurement is only a part of the problem.

It is possible to own a system outright and have it change nothing. Here in Quebec, we have a good example with the Panier Bleu which was supposed to compete with Amazon, but never had adoption by the general public. If the domestic option is slower, or clunkier, or simply put doesn’t meet the user’s needs, people will keep reaching for the foreign one. They will do it without a thought, out of sheer convenience, a thousand times a day. And at that point you still don’t rely on a sovereign infrastructure. You have a sovereign line item on a balance sheet, and a population whose actual behaviour runs through somebody else’s system.

Pix is interesting precisely because nobody adopted it out of patriotism. They adopted it because it was available to all, but mainly because it was better than the alternatives, in ways that are mundane and specific and entirely about design.

You send money using a phone number, an email, or a tax ID. Not an account number, not a routing code. Before you confirm, you see the recipient’s real name, which quietly removes the fear of sending money into a void. It settles in seconds, around the clock, on weekends and holidays. For individuals it is free. For merchants it costs 0.33%, against roughly 1.13% for debit cards and 2.34% for credit. A street vendor can accept a payment by showing a QR code, with no terminal, no contract, and no hardware.

Every item is a design decision: identifier choice, confirmation feedback, latency, fee structure, hardware dependency. Not one of them is a matter of who owns the rails. They are decisions about how the thing feels in the hand of a person trying to get paid for a coconut. Those decisions are why 93% of a country uses it.

The system also cost about four million US dollars to build. That number should be uncomfortable for anyone who assumes sovereign infrastructure is primarily an expenditure problem.

I felt this from the other side when I got home. Coming back to Canada, I wanted to send Interac a set of feature suggestions. Some of what Pix does could reasonably be borrowed. I went looking for a way to reach them and the only channel I could find was a complaint form. So I used it, and I wrote, more or less, that they might take inspiration from what Brazil had done.

I tell that story because the absence of a channel is itself a design fact. An institution with no route for user input is an institution that will be out designed, and being out designed is how you lose the sovereignty you thought you had already bought.

None of this is unique to payments. Substitute cloud infrastructure, health records, identity systems, defence software, or any of the domestic AI initiatives currently being announced. In each case the public conversation is about ownership and the operative question is adoption. If the people who are supposed to use the sovereign system find it worse than the foreign alternative, they will route around it, and the sovereignty will exist on paper.

This is not a new observation for me, only a larger version of one. I wrote recently about an eight-week experiment in getting a group to connect, where the participants all agreed on a communication channel and then most of them never used it. Coordination quietly defaulted to one-to-one messages and the initiative faded. A channel everyone agrees on is not a channel everyone uses. That was true of a small group with a messaging app. It turns out to be just as true of a nation with a payment system, a cloud, or a defence platform. Agreement is procurement. Adoption is design.

I’ll note, because it would be strange not to, that Pix is currently the subject of a United States trade action, and that a decision is expected shortly. I’m not going to argue that dispute here. But there is something worth observing that has nothing to do with who is right. A payment system that nobody used would not be the subject of a trade dispute at all. Adoption is what made the thing real enough to matter, to Brazilians first, and then to everyone else.

Which brings me back to where I started. Sovereignty isn’t secured in the contract, or the press release, or the data centre opening. It’s secured in the moment someone chooses your system over the other one, because yours is better. That moment is a design outcome. If we want sovereign systems, we have to build systems people would choose even if they weren’t sovereign.

Otherwise we continue to be dependent.