Is there a price per employee or per message?

None of the programs we track publishes a per-unit price. They publish ranges per deal and name the factors that move it, such as volume, depth, years of history and the systems connected.

Why does one buyer publish $4M+ and another publish nothing?

Each program decides what to show. Turing Project Lazarus publishes no range. Mercor publishes one per industry listing rather than a single figure. Turing pays the company a share each time it licenses the record to a buyer, with the percentage set in the agreement, so one range would not describe its deals.

How much is my company data worth to AI?

Q: Will a bigger company get a bigger offer?

Not necessarily. Polyshares says the length and texture of the record matter more than headcount or revenue. Headcount is one input among several.

What do AI data buyers publish?

These are the ranges each program shows on its own site. Every figure is as published by that buyer, checked Nov 7, 2026. Offers vary, no buyer guarantees a figure, and we have not verified any of them against signed deals.

Program	Published range	What the buyer says alongside it
micro1	$100K to $3M+, as published by micro1, checked Nov 7, 2026	Tiers of $100K to $500K, $500K to $3M, and $3M+. Agent trajectories, an optional add-on, are described as $1M to $2M a year, recurring
Handshake AI	$100K to $4M+, as published by Handshake AI, checked Nov 7, 2026	Its FAQ says “Payments to date have ranged from $100K to $4M”
Endless Data	Up to $4M, as published by Endless Data, checked Nov 7, 2026	A typical range appears only after you enter a work email in its estimator
Polyshares	$100K to $2M and above, as published by Polyshares, checked Nov 7, 2026	“Typical deals run $100K to $2M and above.”
SimpleClosure AssetHub	Tens to hundreds of thousands of dollars, as published by SimpleClosure, checked Nov 7, 2026	The range is for unique assets
Mercor	$10K to $1.5M across 25 industry listings, as published by Mercor, checked Nov 8, 2026	Each listing has its own range, e.g. law firm $50K to $1.2M; pay also depends on data volume, tools and depth
Turing Project Lazarus	Not published	Pays a revenue share; the percentage is set in the agreement

The buyer comparison keeps these figures current, with the date they were last checked.

Why are the ranges so wide?

Because buyers are pricing different records, from different companies, under different deal structures. The factors they name are consistent enough to be useful.

Volume, tools and depth. Mercor: “Compensation is based on data volume, the tools you connect, and the depth of your data” (Mercor buyer page).

Length and texture over size. Polyshares: “The length and texture of the record matter more than headcount or revenue” (Polyshares buyer page). It lists a short operating history among the things that cap value (source).

History across systems. micro1 says several years of records across multiple business systems are generally a stronger fit, and sets no fixed minimum volume (micro1 buyer page).

Size and history as a starting point. SimpleClosure writes that “Company size and history are a good starting point” (source), and ties its range to unique assets (source).

Where the company operates. micro1 says demand is strongest in the US, UK and Canada (source). Turing names the US, UK, Canada and the EU (source), and Endless Data prioritises the US, UK and Canada (source).

What the company does. micro1 and Turing both say they are strongest in software and engineering, finance and accounting, customer support, sales and CRM, legal and contracts, and operations and logistics (micro1, Turing). Our industry pages cover what each sector can and cannot include.

Scope moves the number too. Every system you leave out, and every category you have to exclude, makes the record smaller. That is usually the right trade; what company data qualifies explains which exclusions are not optional.

What comes off the top before you are paid?

A published range is not what reaches your account if someone takes a share of it. The fees each program publishes for the company side:

Program	Fee to the company, as published
Polyshares	“No fee, no commission.” (source)
Endless Data	“No commission on your side, nothing deducted later.” (source)
Mercor	“no upfront costs, setup fees, or deductions from the payment” (source)
Turing Project Lazarus	Nothing upfront; “Preparation costs are covered from sales, never charged to you.” (source)
Handshake AI	Transfer help at no cost; no fees to the company mentioned (source)
SimpleClosure AssetHub	A percentage only on success; the percentage is not published (source)
micro1	Not published

Turing says preparation costs are “covered from sales”; ask whether that happens before or after your share is calculated.

Why should you never trust a single number?

A range is a claim, not a quote. Most top ends are open: $3M+ at micro1, $4M+ at Handshake AI, “and above” at Polyshares. The figures describe different things: Handshake AI’s FAQ reports what it has paid “to date”, and Polyshares describes “typical deals”. None of them is a price for your company.

Estimators are not offers. micro1 runs a payout estimator, and Endless Data shows its typical range only after you give a work email (source). Offers come after review: Polyshares makes an offer after its team reviews your intake (source), and Handshake AI says it returns pricing in 12 to 24 hours (source).

The amount can change after delivery. micro1 pays “within 30 days of acceptance” (source) and Polyshares pays once the data is approved (source). An offer assumes the data matches what you described. How payouts work covers triggers and timing.

Structures do not compare directly. A fixed sum, a set of milestone payments and a revenue share with no published range cannot be lined up on one axis. Neither can an exclusive license and a non-exclusive one: an exclusive deal costs you every future license of that data, so it should pay more. See exclusivity explained.

How do you get a real number?

Write down what you have: years of history in each system, how many people used it, the share in English, and what must stay out.
Run the value check. It scores fit out of 100 and lists the programs that match. It is an estimate of fit, not a price.
Ask two or three programs that fit for an offer, using the same scope for each. The compare page shows their criteria side by side.
Get every number in writing before data moves. Handshake AI puts amounts in writing before transfer (source), and Endless Data agrees a fixed sum “before anything is shared” (source).
Compare amount, structure, payment trigger, exclusivity and deletion together, using the terms table in how licensing company data to AI works.

Why doesn’t SellData.ai publish its own estimate?

Because it would be a guess dressed up as a number. The only figures on this site are the ones buyers publish, each with the date we checked it. When we introduce you to a program, the buyer makes the offer, and we help you read it next to the others.