Staged payments and earn-outs in a TLD sale

A seller and a buyer agree that a TLD is worth having and disagree on the number. The buyer offers to bridge the gap: part of the price at closing, the rest later, perhaps only if the namespace performs. In the sale of a company or a DNS portfolio this is routine, and both sides have decades of contract practice to lean on.

An onchain TLD makes the same request harder. The asset is a token, it moves from one wallet to another in a single transaction, and nothing in that transaction remembers that money is still owed. Once the buyer holds the token, the seller holds a promise. Once the seller is paid in full, the buyer has no price left to hold back if a problem appears.

This article is for both sides of that table. It explains why the mechanics of a token make deferral awkward, the ways the deferred part can be secured, what an earn-out on a TLD could honestly be measured on, and where each structure leaves the seller or the buyer exposed. It is information in general terms. Contract law, insolvency rules and tax treatment vary by country, and the reader’s own counsel decides.

Three different reasons to defer part of a TLD price

“Pay some of it later” covers three requests that look alike and are not. Each has its own logic, its own winner and its own drafting. Before a structure is discussed, both sides should say which one is on the table, because a negotiation in which the buyer means one and the seller hears another wastes weeks.

Installments are seller credit. The price is fixed and owed in full, and only the calendar is negotiated. A seller who accepts them is lending to the buyer, and should think like a lender: who is the debtor, what does it own, and what happens if it stops paying.

A holdback protects the buyer. Part of the price is kept aside for a period in case a statement the seller made at signing turns out to be wrong, for example about names already issued under the TLD or about the terms attached to it at the registry. If nothing surfaces, the amount is released. The price was never in doubt, only the seller’s access to all of it.

An earn-out is a disagreement about value turned into a formula. The seller believes the namespace will attract registrations. The buyer is not sure and does not want to pay for a forecast. They agree a base price and a further amount that is owed only if a measured result is reached. This is the hardest of the three to write well, and most of this article is about it.

What a one-transaction transfer takes away from the seller

A TLD recorded onchain is a token held in a wallet. Whoever controls the keys of that wallet controls the TLD, and the transfer to a buyer is one transaction. It is final in the practical sense that matters here: no registry clerk, registrar or court officer can press a button that moves the token back. Only the wallet that holds it can send it anywhere.

In the DNS, a seller who is not paid has more levers: a domain sits at a registrar under an account, and a court order addressed to a registrar or registry can be carried out by an intermediary. None of that is automatic or cheap, but the tools exist and lawyers know them. This is one place where the DNS is simply better suited to a slow, staged deal.

With a token, three consequences follow for a deferred price.

There is no lien built into the asset. A contract can say the buyer may not resell the TLD until the price is paid in full. The token does not read the contract. If the buyer transfers it to a third wallet, the seller’s remedy is a claim against the buyer, not the return of the TLD.

Enforcement runs through the person, not the asset. A judgment for the unpaid balance is worth what the debtor is worth, in a place where the judgment can be enforced. A buyer that is a newly formed entity with one asset, held in a wallet whose controllers are not known to the seller, is a weak debtor however sound the contract reads.

Control cannot be split. Issuing names under a TLD is reserved to its holder, and the registry offers no way to delegate management. So there is no halfway position in which the buyer operates the namespace while the seller keeps a hand on the token. Either the buyer holds it and can do everything, or someone else holds it and the buyer can do nothing without that person acting.

Five ways to secure the deferred part, and what each costs

No arrangement removes the risk. Each one moves it to a different party or turns it into a cost.

An unsecured promise

The token moves at closing against the first payment, and the contract records what is still owed. It can be reasonable when the buyer is an established organization with assets, a reputation and a legal address the seller’s counsel can reach. It is rarely reasonable against a special-purpose buyer.

The deferred amount locked at closing

The buyer funds the whole price at closing. The first part goes to the seller with the transfer, and the rest sits in an escrow arrangement that releases it on a date or on a condition. The seller’s credit risk on the buyer disappears, because the money already exists and is out of the buyer’s hands.

The cost is on the buyer’s side: installments that were meant to ease cash timing no longer do so. This structure fits holdbacks and earn-outs far better than budget-driven installments. It also raises a question that must be answered in writing before closing: who decides that the condition is met? If the answer is “both parties agree”, the arrangement needs a tie-break, or the money can stay locked while the two sides argue.

A guarantee from a stronger party

A parent company or another creditworthy party promises to pay if the buyer does not. Its value depends on who the guarantor is and on the law that governs the guarantee, both of which the seller’s counsel should check rather than assume.

The TLD held by an agreed third party

The token goes at closing to a wallet controlled by a party both sides trust, or by a multisignature wallet, which is a wallet that needs several keys to approve a transaction. It moves on to the buyer on full payment and back to the seller on default. The seller is protected. The buyer is not operating: since only the holder can issue names, nothing can be issued unless the third party agrees to act on instructions, which turns a custodian into an operator and brings its own liability questions.

The seller keeps the TLD until paid

The mirror image. The buyer pays in stages and receives the token with the last one. Now the buyer carries the risk: it has paid for an asset that sits in the seller’s wallet, and its remedy if the seller disappears is a claim against a person. If the seller issues names on the buyer’s instructions in the meantime, the deal has become closer to a license with a purchase at the end, which is a different agreement and deserves to be drafted as one.

A worked split: USD 400,000 paid seventy and thirty

The figures in this section are assumptions chosen to make the arithmetic easy to follow. They are not market prices and say nothing about what any TLD is worth.

Assume a seller and a buyer agree a total price of USD 400,000 for a TLD. Assume they agree that 70 percent is paid at closing, against the transfer of the token, and 30 percent is deferred. The payment at closing is USD 280,000. The deferred amount is USD 120,000.

Paid at closing, 70 percentUSD 280,000

Deferred, 30 percentUSD 120,000

Both amounts follow from the assumed price and the assumed split. Taxes, adviser costs and the time value of money are not included.

The same two amounts mean very different things depending on which of the three structures applies.

If the USD 120,000 is an installment, the seller has sold for USD 400,000 and lent USD 120,000 of it to the buyer. The seller’s question is whether the certain USD 280,000 would be an acceptable outcome if the rest were never paid. If it would not, the structure needs security from the previous section, or the split needs to change.

If the USD 120,000 is a holdback, the seller expects to receive all of it and the buyer expects to keep some only if a statement proves wrong. The negotiation is about the period, the kinds of claim that count, and whether the buyer may withhold the whole amount or only the amount of the claim.

If the USD 120,000 is an earn-out, the honest way to read the deal is that the price is USD 280,000 and the seller has a chance at up to USD 120,000 more. Suppose, as a further assumption, that the earn-out is proportional to a registrations target and that half of the target is reached within the period. Half of USD 120,000 is USD 60,000, and the seller’s total is USD 340,000. If the target is reached in full the total is USD 400,000, and if no qualifying registration occurs it stays at USD 280,000.

A seller comparing this with an all-cash offer should compare it with a figure between USD 280,000 and USD 400,000, and not with the headline.

What an earn-out on a TLD can be measured on

An earn-out is only as good as its measure.

The registry’s record is public, so the names issued under a TLD can be looked up on a block explorer by anyone. A count of names registered under .yoursector between two dates does not depend on the buyer’s bookkeeping. Both parties, and any third person asked to settle a disagreement, can read the same data and should reach the same number. That makes registrations the natural measure for an earn-out on a TLD.

A count of registrations is objective, and it is also crude. It treats a name issued for free and a name sold at a high price as the same event. The holder of a TLD sets the price of the names under it and may price them differently, so a count says little about income unless the contract fixes what qualifies.

A measure based on money is closer to what the seller wants to share in, and further from the record. Where the TLD takes part in the registry’s royalty program, the holder receives a share of each registration fee, subject to the registry’s terms and activation conditions, and the program carries an annual cost of its own that any income-based formula must account for. The share applies to registrations, so an earn-out should not be drafted on the assumption that renewals feed it. A formula built on income will rely on statements the buyer obtains from the registry, which reintroduces the trust problem the public record had removed. The site’s subname revenue calculator can help both sides test what a given count might mean in income before they argue about a target, as an estimate under their own assumptions.

A sound compromise is a count with qualifying conditions: names registered to wallets that are not controlled by either party, at or above a stated minimum price, within the period.

How each side can bend a registrations count

A public number is not a neutral number. Both sides should read this section as a list of behaviors to recognize and write out of the deal.

The buyer controls issuance. Only the holder issues names, so the buyer decides when registrations open, to whom, and at what price. A buyer who delays opening until the day after the earn-out period ends has broken no rule of the registry. Whether it has broken the contract depends on whether the contract says anything about it.

The buyer controls price. A high price suppresses the count. A change of plan, such as issuing names only to the buyer’s own group, can reduce third-party registrations to zero for sound business reasons that have nothing to do with the seller. An organization that acquires .yourbrand to use names such as pay.yourbrand internally may never open registration to the public at all. An earn-out on registrations is simply the wrong tool for that buyer.

The seller, or anyone friendly to it, can register. If names are open to the public and the measure is a bare count, registrations made near the end of the period by wallets connected to the seller would lift the number. The cost of those registrations may be small next to the deferred amount. Distinct wallets are no cure, since creating wallets costs almost nothing.

Neither side controls the surroundings. Onchain names are not part of the ICANN DNS root; wallets and browsers support them only in part, and that coverage changes. A wallet adding or dropping support, or a change in the registry’s own terms, can move registrations more than any effort by the buyer. An earn-out makes the seller a silent partner in those risks without any say in how they are handled.

Clauses that keep a staged TLD price enforceable

The drafting belongs to counsel, and the right words depend on the governing law. The points below are the commercial decisions the lawyers will ask the principals to make.

The measure, defined to the last detail. What counts as a registration, on which registry record, between which two moments, read by whom. State whether free names, names issued to the buyer’s group, names issued to the seller or its associates, and names later given up are counted.

The period. Long enough for the buyer to launch, short enough that the seller is not tied to events it cannot influence. If the buyer needs time before opening registrations, a period that starts at opening, with a long-stop date after which it starts anyway, is fairer than one that starts at closing.

The buyer’s conduct. The seller will ask for commitments: to open registrations by a date, not to reprice in order to defeat the earn-out, not to move the TLD to another wallet without the obligation following it. The buyer will resist anything that tells it how to run its own namespace. A common landing place is a short list of things the buyer will not do for the purpose of reducing the payment, plus an acceleration clause.

Acceleration. Events that make the whole deferred amount due at once: a resale or transfer of the TLD, the buyer’s insolvency, or a decision to close the namespace to third parties. Acceleration matters most with a token, because the asset can leave the buyer’s wallet in a moment.

Floor, cap and shape. Whether the earn-out is all-or-nothing at a threshold, proportional, or stepped. All-or-nothing formulas invite disputes when the result lands just under the line.

Who decides, and how fast. A named independent person who reads the public record and issues a figure, with a short timetable and a rule on who pays for it. Without this, a locked amount can stay locked.

The seller’s own statements. A holdback only works if the statements it backs are precise: who controls the wallet, what names the seller holds under the TLD and whether they transfer with it by agreement, what third parties hold, and on what terms the TLD sits at the registry. If the TLD sits under earlier royalty terms, ask the registry before closing whether a buyer keeps them; do not assume it and then test it through a holdback.

When one payment beats a staged price

Staging has a price of its own. It adds drafting, a period of shared exposure, and a relationship between two parties who may have preferred to part at closing.

A staged price tends to earn its place in a narrow set of cases. The buyer is creditworthy and its reason for deferring is timing. Or the gap between the two valuations rests on one identifiable question, such as whether third parties will register names, and both sides accept that the answer will be visible in the public record within a reasonable period. Or the buyer has a specific, limited concern that a holdback covers more cheaply than a price cut.

Outside those cases, the alternatives are usually better. A lower all-cash price converts the seller’s uncertain upside into certainty and gives the buyer a clean asset. A longer period before closing, during which the buyer completes its internal approvals, removes the need for installments without separating payment from transfer.

It is also fair to ask whether the gap is real. Parties sometimes reach for an earn-out because neither has done the work of valuing the TLD on evidence. A valuation built on stated assumptions narrows many gaps that a formula would only postpone.

Preparing a staged structure before the term sheet

A staged price is easier to negotiate when the facts under it are already assembled. The TLD readiness score is a quick way for a seller to see what is missing, and so what a buyer might ask a holdback to cover.

1. Name the reasonAsk the other side whether the deferral is about cash timing, cover for the seller's statements, or a different view of future registrations. Write the answer into the term sheet in one sentence.
2. Fix the closing amount firstAgree the amount paid against the transfer before discussing the rest. The seller decides whether that amount would be acceptable alone, and the buyer decides whether it can fund it.
3. Choose where the deferred amount sitsDecide between an unsecured promise, an amount locked at closing, a guarantee, or a holding arrangement for the token, and record who bears the cost of the choice.
4. Define the measure and test itFor an earn-out, write the definition of a qualifying registration and have each side compute the count from the public record of a past period. Two different answers mean the definition is not finished.
5. Agree the decider and the timetableName the independent person who settles a disagreement, the days allowed for each stage, and the events that make the full amount due at once.
6. Hand the structure to counselEach side's lawyer and accountant review the governing law, enforcement against the debtor, and the tax treatment of deferred and contingent amounts before anything is signed.

Closing itself keeps one rule whatever the structure: the token and the payment due at closing change hands together. Hush Partners runs closings through onchain escrow on an aftermarket, so that neither side has to move first for that part of the price. Anything deferred beyond that moment rests on the contract and on the security the parties chose.

A holder can mandate the desk through a sell-side mandate, and an organization through a buy-side mandate; the stages of a file run from mandate to transfer, and a standard mutual NDA applies to each exchange. If you are weighing an offer that includes a deferred part, or preparing to make one, the practical next step is to write down the reason for the deferral and the amount you would accept at closing alone, then take both to your counsel. If you want the desk involved, the contact form is the only way in.

Questions people ask

Can a smart contract return the TLD to the seller automatically if the buyer does not pay?

Only if the token sits in that contract or in a wallet bound by it, and not in the buyer’s own wallet. In that case the buyer does not hold the TLD during the period and cannot issue names under it. Ask counsel and a technical reviewer before relying on one.

What happens to an earn-out if the buyer sells the TLD on before the period ends?

Whatever the contract says, and nothing if it is silent. The new holder is not a party to the original agreement and the token carries no trace of it. This is why an acceleration clause on transfer, making the full deferred amount due at once, is one of the first things a seller’s lawyer asks for.

Is an earn-out a sign that the buyer doubts the TLD?

Not necessarily. It often means the buyer’s approval process cannot pay for a forecast, or that the buyer’s planned use differs from the seller’s picture of it. Asking which of the three reasons applies usually tells the seller more than the proposed formula does.