Stablecoins, chains and the cost of getting paid.
Working notes from a company in the middle of building this. Written mostly to force ourselves to be specific, and published because the reasoning is more useful to you than the conclusion.
Notes, positions and things we could be wrong about
Your business became the integration layer
Eleven steps between “the customer paid” and “the money is usable”, and why none of them are your job.
Why we will never write a bridge
Burn-and-mint against lock-and-wrap, what a wrapped token actually is, and why a contract we authored holding your funds is custody whatever we call it.
Three rates, and only one of them moves money
Reference, executable, customer price. What each one is for, why bundling them is the industry’s favourite trick, and why we will not quote a locked rate.
The number that decides whether this company exists
What share of inbound payment volume arrives through a rail the receiver cannot use — and why counting how often it happens would mislead us.
Unknown is a state. It is not a failure.
The most dangerous moment in a cross-chain payment is the one where nothing has happened yet, and the honest engineering response is to do nothing at all.
Dates are when each piece was written. Everything here reflects what we believed at the time; where we later find out we were wrong, we correct it in public rather than quietly.
Tell us which of these is wrong.
Especially the market-size one. If you run a Nigerian business paid in stablecoins, you know something we are currently guessing at.