The trust stack

The word stack has two parents. One is software, where it denotes the chain of libraries and runtimes a program depends on. The other is finance, where it denotes the chain of institutions a trade depends on — clearinghouse to broker to custodian to bank. The interesting feature of a crypto swap is that you are picking both stacks at once, usually without being told you are doing it, and the financial one matters more.

The traditional stack, briefly

A buy-side equity trade in 2026 is touched by, at minimum, an executing broker, a prime broker, a clearinghouse, a depository, and a custodian. Each of these institutions is regulated by a different body, audited on a different cadence, and capitalized to a different standard. The system works not because any single node is impervious but because the failure of one node is meant to be absorbed by the others. The DTCC's entire job description is "be the thing that does not fail when something else does." The stack was engineered, over forty years, to make cascading failure expensive and contained.

The crypto stack, by accident

The equivalent stack in crypto was not engineered; it accreted. A user moving Bitcoin into a yield-bearing DeFi position is depending on: a centralized custodian (who holds the actual BTC), a merchant (who minted the wrapped version), the EVM chain hosting that wrapped version, the smart contract instantiating the token, a DEX router, a lending protocol, an oracle providing prices to the lender, a governance system controlling the oracle, and a bridge if anything crosses chains. That is nine counterparties. None of them has the contractual obligation to make the user whole if a peer fails. Most of them have an explicit disclaimer to the contrary.

The user is told they are doing "DeFi." What they are doing is consenting, all at once, to a stack that TradFi spent decades trying to architect away.

Intent settlement is stack compression

The case for intent settlement is most legible if you stop thinking about it as a swap mechanism and start thinking about it as a deletion of trust dependencies. The traditional cross-chain swap stacks: custodian + merchant + bridge + DEX + (sometimes) wrapper. Intent settlement replaces that with: protocol atomicity guarantee + solver collateral economics. Two dependencies, not five. Each of the remaining two is enforceable on-chain in real time rather than legible only through audits, attestations, and retrospective post-mortems.

This is the same move TradFi made in the 1970s and 1980s when it consolidated dozens of regional clearinghouses into the DTCC. Fewer trust nodes, more capital behind each one, clearer rules for what happens when one fails. The difference is that crypto is doing it without the forty-year regulatory runway, which is faster but also riskier, which is in turn the whole story of crypto.

What a swap UI is for

A swap UI's job is not to look pretty. Its job is to be honest about the stack the user is consenting to. Most swap UIs are not honest, in two senses: they hide the stack (because surfacing it would be discouraging), and they pretend the stack is shorter than it is (because the team building the UI does not always know what it is either).

Flip is built on a short stack. That is the whole product argument, restated as engineering: the interface does not get to be longer than the system underneath it.