Upfront payment or revenue share?

An upfront payment fixes your price and leaves the risk of reselling your data with the buyer; a revenue share pays you only as your data is licensed onward, so you carry that risk in return for a share of every later sale. Which suits you depends on when you need the money, whether the license is exclusive, and whether you can check the reporting.

How do the two models divide risk?

	Upfront payment	Revenue share	Hybrid
What you receive	A fixed amount for the license	A percentage of what buyers pay for your data	A fixed amount, then a percentage of later licenses
When you are paid	On a set event: approval, delivery or acceptance	When a sale happens and the buyer pays	Fixed part on a set event; the share as sales happen
Who carries resale risk	The buyer	You	Shared
Upside if the data sells many times	Stays with the buyer	Shared with you	Shared with you

An upfront buyer pays before it knows how often it will resell your data. Sunset, for example, says it buys a closing startup’s dataset directly and will “assume responsibility for licensing it to downstream partners”. If the data never resells, the buyer absorbs that. If it resells many times, the buyer keeps the gain. Its price reflects both possibilities.

A revenue share reverses this. You are paid when someone else makes a sale. For Troveo’s content-owner program, its support article says payouts follow licensing, invoicing of the AI client and receipt of the client’s payment, with “no fixed payout schedule”; Troveo has not published whether business data follows the same path (checked October 2026).

Does “upfront” mean paid when I sign?

Usually not. In the programs Data Licensing Report covers, “upfront” means a fixed amount that does not depend on resale, and it is still paid on a later event such as approval or delivery. Each named program’s trigger is compared in how data licensing offers work.

How does exclusivity change the choice?

Some programs tie the payment model to exclusivity. License My Data states the link directly: an “exclusive buyout pays once, upfront”, while a non-exclusive license “can be placed with several buyers and paid on each.” Every case study Polyshares publishes is a single sum for a “perpetual exclusive license”.

A revenue share combined with exclusivity makes the provider the only channel selling that data for the exclusive period. Troveo says about 95 percent of its licensors have signed exclusively, across its programs, and that business-data exclusivity is “time-limited and it has to be earned”, with exit terms in the agreement. Under that structure, the length of the exclusive period and the exit terms decide how long your share depends on one seller.

How does Replay’s hybrid work?

Replay says its offers are “typically structured as an upfront cash payment plus a perpetual revenue share”. A hybrid gives you a floor and keeps part of the upside.

As of October 2026, Replay does not publish when the upfront part is paid, the share percentage, how often the share is paid, or how “perpetual” is defined. Its published value tiers, such as $100K to $1M for companies of 50 to 250 employees, do not say whether they cover the upfront part alone or include expected revenue share (Replay). Those are the questions to ask of any hybrid.

Staged payment is different: Corpus offers larger engagements “upfront plus delivery tranche”, two fixed payments with no share.

What should I ask about a revenue share?

As of October 2026, none of Troveo, Replay or Defined.ai publishes the percentage it shares, and Scale AI publishes no split for the recurring revenue it describes. The one published split Data Licensing Report found is on HUD’s DataVendor, a marketplace mostly for codebases, where HUD says the seller “receives 80% of the final payout, with 20% covering grading and packaging”.

Get these in writing:

• Percentage. Your share, and whether it is the same for the first sale and every later one.
• Basis. Gross (what the buyer pays) or net of the provider’s costs, and if net, which costs. Troveo says it takes “no deductions from your payouts”; the share those payouts represent is not published.
• Unit. Per sale, per license, or per year of a license. A buyer that renews annually may produce a payment each year or only once.
• Reporting. A statement per sale showing the amount received and your share, and a right to audit.
• Timing. How long after the buyer pays you are paid.
• Term. How long the share lasts, and what happens to it if the license or the exclusivity ends.

Which named programs use which model?

As of October 2026:

Program	Model	What it says
Corpus	Upfront	One-time payment, “no revenue share” (source)
Mercor	Upfront	Pays directly; no fees (source)
Polyshares	Upfront	Pays directly; no seller fee (source)
Sunset	Upfront	Buys closing startups’ datasets outright (source)
License My Data	Either	Exclusive buyout paid once, or non-exclusive paid per placement (source)
Replay	Hybrid	Upfront cash plus a perpetual revenue share (source)
Defined.ai	Revenue share	Licensing fees from marketplace sales (source)
Scale AI	Recurring	Opportunity for recurring revenue; no split published (source)
Troveo	Revenue share	A share of every sale; no percentage published (source)

The full lists are providers that pay upfront and providers that share revenue.

Which should I choose?

• If you need the money by a known date, upfront fits: the amount and the trigger are fixed.
• If your company is closing, a share that pays over years can outlast the entity owed it. Troveo advises setting up licensing before the entity dissolves (Troveo).
• If the license is non-exclusive and you can wait, a share can pay more than once.
• If a share comes with exclusivity, weigh it against upfront offers over the length of the exclusive period.
• If you are offered a hybrid, decide whether the upfront part alone is acceptable, then treat the share as upside.

Troveo’s business-data page shows an illustrative payout of $850,000 across four sales; what a share pays depends on a percentage that is not published and on how many sales happen. The guide covers the other terms.