How AI data buyers' payout estimators work

Three of the AI companies we list publish a payout estimator for company records: micro1, Telegraph and Grepped. Each takes two or three answers about the company and returns a dollar range. The formulas below are the ones their pages run, copied from each site and checked against the live estimators on October 26, 2026. The company worth estimator runs all three on one set of answers.

All three reward headcount and years, and two reward a US address

Every estimator grows with the number of people and the number of years the company has run, because more people over more years means more records. micro1 and Telegraph multiply the result by where the company is based: the United States counts in full, Canada and Europe at three quarters, and everywhere else at 30%. Neither lists the United Kingdom, so a UK company picks Europe. Grepped asks about revenue instead of location.

micro1 starts at $105,000 and adds people, years, and people times years

micro1’s page sends the answers to its server, so the formula was fitted from the server’s replies to 84 sample companies. For a company with p employees and y years of data:

• Low end: ($105,000 + $1,749.75 × p + $8,750.25 × y + $11.81 × p × y) × location
• High end: ($175,000 + $2,916.25 × p + $14,583.75 × y + $275.63 × p × y) × location

Employees run from 1 to 10,000 and years from 1 to 50, and the result is shown to the nearest $1,000. The fitted formula lands within $25 of micro1’s figure for companies with 40 or more people or five or more years of data, and up to about $3,700 off for a very small, young company.

The last term is the one people miss. For a 100-person company with 10 years of records it adds $11,812 to the low end and $275,625 to the high end. For a 1,000-person company with 30 years of records it is most of the figure: $354,000 of the $2.5M low end and $8.3M of the $11.8M high end. Leave it out and a large, old company is undercounted by more than half.

People	Years	micro1’s range (US)
30	5	$203K to $377K
100	10	$379K to $888K
500	20	$1.3M to $4.7M
1,000	30	$2.5M to $11.8M

micro1 labels the result an indicative estimate. Its published tiers are $100K to $500K for qualified partnerships, $500K to $2M for large-scale datasets or ongoing participation across teams, and $2M+ for highly unique proprietary data, and it says it pays in full within 30 days of acceptance.

Telegraph reads a four-row table, adds a per-year slope and floors the result at $100,000

Telegraph’s estimator is a table in its page script. Employees are held to 20 to 200 and years to 3 to 20:

People	Low at 10 years	High at 10 years	Per year, low	Per year, high
20	$163,000	$439,000	$9,000	$20,300
50	$289,000	$611,000	$9,430	$28,650
100	$383,000	$897,000	$10,030	$42,570
200	$572,000	$1,470,000	$11,220	$70,400

A headcount between two rows is interpolated. Each year above ten adds the per-year amount, each year below takes it away, the low end never drops under $100,000, the location factor is applied, and the result is rounded to $5,000. A 60-person US firm with 5 years of records gets $260,000 to $510,000. Telegraph never calls the figure indicative; every deal is “subject to data review and agreed terms”.

The table lines up with the three deals Telegraph lists: $350,000 to a 60-person accounting firm, $450,000 to a 70-person engineering firm and $650,000 to a 100-person freight broker, each inside the range its estimator gives for that headcount at ten years.

Grepped multiplies a revenue base by a team-size curve and an age curve

Grepped’s business estimate starts from annual revenue: a base of $12,000 for a pre-revenue company, $20,000 up to $1M, $29,000 for $1M to $10M, $55,000 for $10M to $50M, $95,000 for $50M to $200M, and $150,000 above that. A company that prefers not to say gets the $29,000 base.

The base is multiplied by a team-size curve (0.8 at one person, 1 at 24, 1.3 at 99, 1.8 at 1,000) and an age curve (0.4 for a new company, 1 at ten years, 1.6 at fifty), each interpolated between points. That gives the low end, rounded to $500, and the high end is 1.43 times the low. A 30-person company that is 5 years old with no stated revenue gets $24,000 to $34,000. Grepped’s own quote form asks more, including which systems the company holds, and prices from that.

None of them ask what survived, how clean it is, or who can sign

The three estimators read size, age and location or revenue, and no more. They don’t ask which systems the records sit in, how many years of history are still in them, how much personal data they carry, or whether someone has the authority to license them. That is why each range is an estimate and the deal is priced after a review of a sample. The company fit check asks those questions and shows which of the listed companies take what you have.

The other listed companies publish a range instead

• Scale AI: $10K to $1M+ per partnership, scaling with cadence, its own illustration.
• Mode: operating history “could be worth $100K-$5M”, its own claim.
• Turing: a single payout of up to $1M for a company’s operating record, running, winding down or closed.
• Miro Advisory: $100K to $4M+ for an operating dataset, indicative.

Each of those is the company’s own figure and not a quote. A closed company should note that micro1’s estimator is built for operating companies and its program for closed companies publishes no figure; Telegraph’s and Grepped’s estimators don’t ask whether the company is still open.