Escrow purchase or multi-leg barter: where the 60-day lock decides

A premium domain deal is a distributed transaction with no rollback: once a registrar accepts a push or an auth code, the asset has moved and nothing in the protocol pulls it back. Direct purchase through escrow keeps that transaction to a single leg with one release condition. Barter adds legs, and every added leg is another place where one side completes while the other stalls. A founder securing a deep tech name before launch needs to know which design still closes when a participant, a registrar or an asset fails mid-flight.

Two shapes of the same deal

In the escrow model, cash sits with a neutral agent, the seller moves the domain, the buyer confirms control, and only then are funds released. The failure surface is small: if the transfer stalls, the money never leaves escrow and the deal unwinds cleanly. Barter replaces cash with assets that have their own custody chains, whether another domain, a block of equity or a services credit, so each leg needs its own proof of delivery. When two domains are swapped, both must reach a state the agent can verify before either side is released, which means the agent is coordinating two registrars it does not control. The real architecture choice lies in that coordination. The price matters less.

The failure that matters

The characteristic failure is the half-completed swap: your name has already moved to the counterparty, their asset sits frozen at a registrar, caught in a chargeback dispute or simply unverifiable, and you now hold a claim instead of a domain.

Where the decision flips

The boundary is set by the inter-registrar transfer rules in ICANN policy, together with the EPP status codes that registries enforce. After a change of registrant, most registrars still apply a 60-day lock on transfers to another registrar unless the registrant opted out beforehand. If every barter leg can be completed and verified inside the escrow inspection window, the swap behaves like two clean purchases and barter is viable. It can even win, because nobody has to argue over a cash valuation. Once any leg involves an asset that changed hands recently and is still under that lock, or a registrar that cannot push between accounts internally, the legs can no longer settle in the same window. At that boundary direct purchase wins, and the rule is simple: barter only when every leg can reach verifiable delivery within one escrow window, and otherwise split the deal into cash-settled legs.

Keeping the name resolving while custody moves

Whichever path you take, the domain must keep answering DNS queries while ownership changes, and that failure domain is separate from the deal itself. Host authoritative DNS with a provider that depends on neither the selling registrar nor the receiving one, so a stalled transfer or a suspended account never takes down mail or the landing page. Change nameservers before the transfer or long after it, never during. My estimate is that most outages around premium acquisitions come from nameservers changed mid-transfer, not from the transfer itself. The cash path still has a cost in liquidity: a pre-seed company that pays cash for its name spends runway a barter would have preserved, and that trade-off has to be weighed before the design freeze, not after.

Can a two-domain swap be made atomic?

Not at protocol level, because each registry processes its own transfer independently. The closest you can get is a single escrow agent that holds both auth codes and releases nothing until both domains show the receiving accounts as registrant, and that only works when neither name is under a transfer lock.

Back to Deep Barter · Domains

← Back to Deep Barter
Engineering & Architecture Brief

Escrow purchase or multi-leg barter: where the 60-day lock decides

A premium domain deal is a distributed transaction with no rollback: once a registrar accepts a push or an auth code, the asset has moved and nothing in the protocol pulls it back. Direct purchase through escrow keeps that transaction to a single leg with one release condition. Barter adds legs, and every added leg is another place where one side completes while the other stalls. A founder securing a deep tech name before launch needs to know which design still closes when a participant, a registrar or an asset fails mid-flight.

Two shapes of the same deal

In the escrow model, cash sits with a neutral agent, the seller moves the domain, the buyer confirms control, and only then are funds released. The failure surface is small: if the transfer stalls, the money never leaves escrow and the deal unwinds cleanly. Barter replaces cash with assets that have their own custody chains, whether another domain, a block of equity or a services credit, so each leg needs its own proof of delivery. When two domains are swapped, both must reach a state the agent can verify before either side is released, which means the agent is coordinating two registrars it does not control. The real architecture choice lies in that coordination. The price matters less.

The failure that matters

The characteristic failure is the half-completed swap: your name has already moved to the counterparty, their asset sits frozen at a registrar, caught in a chargeback dispute or simply unverifiable, and you now hold a claim instead of a domain.

Where the decision flips

The boundary is set by the inter-registrar transfer rules in ICANN policy, together with the EPP status codes that registries enforce. After a change of registrant, most registrars still apply a 60-day lock on transfers to another registrar unless the registrant opted out beforehand. If every barter leg can be completed and verified inside the escrow inspection window, the swap behaves like two clean purchases and barter is viable. It can even win, because nobody has to argue over a cash valuation. Once any leg involves an asset that changed hands recently and is still under that lock, or a registrar that cannot push between accounts internally, the legs can no longer settle in the same window. At that boundary direct purchase wins, and the rule is simple: barter only when every leg can reach verifiable delivery within one escrow window, and otherwise split the deal into cash-settled legs.

Keeping the name resolving while custody moves

Whichever path you take, the domain must keep answering DNS queries while ownership changes, and that failure domain is separate from the deal itself. Host authoritative DNS with a provider that depends on neither the selling registrar nor the receiving one, so a stalled transfer or a suspended account never takes down mail or the landing page. Change nameservers before the transfer or long after it, never during. My estimate is that most outages around premium acquisitions come from nameservers changed mid-transfer, not from the transfer itself. The cash path still has a cost in liquidity: a pre-seed company that pays cash for its name spends runway a barter would have preserved, and that trade-off has to be weighed before the design freeze, not after.

Technical FAQ: Can a two-domain swap be made atomic?

Not at protocol level, because each registry processes its own transfer independently. The closest you can get is a single escrow agent that holds both auth codes and releases nothing until both domains show the receiving accounts as registrant, and that only works when neither name is under a transfer lock.