How do exchangers settle cross-chain trades without holding your funds in a pooled account

05 September 2026

They do not settle trades at all in the usual sense. Instead, they orchestrate a sequence of independent on-chain transactions, each finalised on its own blockchain, and rely on the user to complete the first leg before the second leg is executed.

The key mechanism is a two-step atomic swap variant adapted for custodial intermediation. You send your tokens to an address the exchanger controls on the source chain. The exchanger waits for sufficient confirmations, then sends the corresponding tokens to your target address on the destination chain from its own operational wallet. No pooled account holds both sides of the trade simultaneously. The exchanger’s liquidity is spread across separate wallets on each chain it supports.

Why this works without a pooled account comes down to timing and risk management. The exchanger does not need to hold your funds in a central pot because it never takes possession of both assets at once. It receives your deposit, confirms it, and then releases its own funds from a different wallet. The gap between receipt and release is measured in minutes, not days. During that window, the exchanger’s balance on the source chain increases (your deposit) while its balance on the destination chain decreases (the payout). The net effect is a shift of liquidity from one chain to another.

This model avoids several problems that pooled accounts create. A pooled account on a single chain is a honeypot for attackers. It also requires the exchanger to maintain a single giant wallet with complex permission controls. Spreading liquidity across per-chain wallets reduces the blast radius of a compromise. If an attacker drains the source-chain wallet, the destination-chain wallets remain untouched.

The exchanger still needs to manage its overall liquidity position. If many users swap from Ethereum to Solana in a short period, the Solana wallet will drain faster than the Ethereum wallet fills. The exchanger must periodically rebalance by buying tokens on the open market or using its own reserves. This rebalancing happens outside the user-facing swap flow and is invisible to customers.

Some exchangers use a variant where they generate a fresh deposit address for each trade, tied to a specific order. This address belongs to the exchanger, but it is not a pooled account in the traditional sense because it is ephemeral. Once the trade completes, that address is retired. This adds privacy and reduces the risk of address poisoning, but the settlement logic is identical: receive on chain A, confirm, send from chain B.

The critical point for users is that the exchanger acts as a trusted intermediary during the settlement window. You must send first and trust that the exchanger will honour the second leg. Reputable exchangers enforce this with automated systems that monitor incoming transactions and trigger outgoing payments programmatically. If the system fails - due to a bug, network congestion, or deliberate fraud - you have no on-chain recourse. The exchange is custodial for those few minutes.

For a deeper look at what actually moves between blockchains during these swaps, see the sibling page "What actually moves between blockchains when you do a cross-chain swap without a bridge". It explains why no tokens physically cross a chain boundary.

The hub page "Swapping crypto across chains" ties all these mechanisms together and shows how the exchanger’s role fits into the broader landscape of moving value between networks without bridges or centralised accounts.

Back to staking