The real cost of building instant payments

Every team evaluating instant payments starts at the same place: a gateway and its per-transaction rate. That number is easy to find, easy to compare, and almost entirely beside the point.

The cheapest part of a payments program is executing the payment. Between “we can push a payment to a card” and “we operate a compliant payments program at scale,” there are at least a dozen capabilities that someone has to own: a sponsor bank and FBO structure, access to each payment rail, KYC and identity, capability and velocity controls, AML monitoring, card validation, settlement and ledgering, treasury and funding, dispute management, an operations surface, and the event infrastructure that ties it all together. A gateway supplies almost none of them.

So the real comparison is not gateway fee versus platform fee. It is one platform fee versus the full cost of assembling and maintaining that dozen yourself: eight or more vendor contracts, a compliance program and the officer to run it, twelve to eighteen months of engineering to integrate systems that were never designed to work together, and the recurring cost of keeping them working as each vendor changes its API. All in, that path runs well past $500,000 in the first year, before a single payment moves at scale.

This paper does that pricing. It walks through each capability a payments program requires, what it costs to buy or build on the open market, and where the hidden costs of integration, float, and opportunity actually sit. It uses public market figures where they exist and labels every illustrative range as such.

Astra is the Payments Cloud: those dozen capabilities, delivered through one platform, under one contract, for a predictable all-in cost. The rest of this paper is the itemized bill for building them yourself.

A gateway moves a transaction. It does not run a payments business.