Free tool and price guide
What is your company data worth to AI companies?
On this page
The short answer
For an established business that qualifies, AI data buyers advertise starting points around $100,000. micro1 advertises $100K+, $500K+ and $1M+ tiers, Handshake AI advertises $100K to $4M+, and Mode advertises $100K to $5M. Reported sales by shut-down startups have been smaller, mostly tens of thousands of dollars.
What sets your number: years of history, how many systems you can include, how well your records connect, and what has to stay out. Use the checker below to see which programs you qualify for today.
Check which buyers your records qualify for
Enter a few facts about your business. The checker compares them with each program's published eligibility and shows what that program advertises paying. It does not invent a price; nobody can price records honestly before seeing what you can license.
What AI data buyers advertise paying
These are each program's own published figures, checked against its site. Several programs publish no range at all, including Mercor, which prices on volume, tools connected and depth (source: Mercor, Monetize your data).
| Program | Advertised pay to the business |
|---|---|
| micro1 | Advertised tiers of $100K+ for a qualified partnership, $500K+ for large or multi-team datasets, and $1M+ for highly unique data. The referral page says $100K to $2M+ for approved data packages. (source: micro1, data partnerships, micro1, company referral) |
| Handshake AI | "Partnerships can range from $100K to $4M+ depending on the data shared." Its estimator example for 100 employees and 10 years in the US shows $383,083 to $896,969 and says it is not an offer. (source: Handshake AI, data partnerships) |
| Nyne | "Commonly $100K or more." Nyne says it will beat a qualifying written offer from another buyer; estimates and "up to" quotes don't qualify. (source: Nyne, sell company data, Nyne, compare data licensing offers) |
| Mode (BizBuyData) | $100K to $5M advertised. Its calculator is labeled indicative only. (source: Mode, data program) |
| Pewter | An estimator example of $385K to $895K for a 100-person, 10-year firm, labeled an illustration based on its partner's payout model. (source: Pewter) |
| HUMXN Research Labs | Paid per license. The release describes payouts of "many times" six figures for some companies. (source: Yahoo Finance, Oct 13, 2026) |
| Endless | "Get up to $4M." A fixed sum is agreed before any data is shared. (source: Endless) |
| Troveo | Revenue share on every sale. "Typical full-company deals start at six figures." It reports $20M+ paid to licensors across all categories. (source: Troveo, business data) |
| Turing (Project Lazarus) | Up to $1M in one payment; standalone codebases $10K to $100K+. (source: aitrainer.work summary of Project Lazarus) |
| Polyshares | "$100K to $2M+." It says deals typically land between $100K and $2M and run higher for large or long records. Its site shows an example offer card for $640,000. (source: Polyshares) |
| SimpleClosure (AssetHub) | "Tens to hundreds of thousands of dollars for unique assets." MIT Technology Review reports 150+ deals since April 2026 and $10M+ expected to be returned by year-end. (source: SimpleClosure AssetHub, MIT Technology Review, Oct 13, 2026) |
| Sunset | No payouts published. Sunset buys the dataset, so it sets the price. (source: Sunset blog, Apr 14, 2026, MIT Technology Review, Oct 13, 2026) |
| DataSupply.ai | No guarantee published. (source: DataSupply, how it works) |
| SHFA.ai | Says medical facilities have been paid $300,000 to $3,000,000; upfront, royalties, or both. (source: SHFA.ai) |
| Avelence | Its estimator example shows $160K to $500K. (source: Avelence) |
Company data prices reported so far
Advertised ranges are marketing. Reported deals are rarer and mostly involve companies that shut down, because their estates can sell without clients or staff to answer to. Here is every public price point we could verify:
| Deal or data point | Price | When | Source |
|---|---|---|---|
| Spirit Airlines records (bankrupt airline) | $10M | Aug 2026 | (source: MIT Technology Review, Oct 13, 2026, The Next Web, on the Spirit data sale) |
| micro1's competing bid for the same records | $12.5M | Sep 2026 | (source: Fortune, Sep 14, 2026) |
| SimpleClosure shutdown sales, 150+ deals since April 2026 | $10M+ expected returned to sellers by year-end | Oct 2026 | (source: MIT Technology Review, Oct 13, 2026) |
| One SimpleClosure example: codebase, plus workspace data | About $25K, plus about $15K pending | Oct 2026 | (source: MIT Technology Review, Oct 13, 2026) |
| cielo24's full Slack, email and Jira archive | Hundreds of thousands of dollars | Apr 2026 | (source: Forbes, Apr 16, 2026) |
| Bulk text licensed for training | About $0.001 per word | 2024 | (source: Reuters, via KFGO, Apr 2024) |
Two patterns stand out. First, scale and uniqueness command the big numbers: Spirit's archive drew competing bids from Google and micro1 (source: Fortune, Sep 14, 2026). Second, the shutdown market is high volume at lower prices. SimpleClosure expects to return more than $10 million across 150+ deals this year, which works out to roughly $65,000 per deal if spread evenly (our arithmetic, from the totals MIT Technology Review reported) (source: MIT Technology Review, Oct 13, 2026).
One useful floor: bulk text has been licensed for training at around $0.001 per word (source: Reuters, via KFGO, Apr 2024). At that rate, a 20 million word archive is worth about $20,000 as raw text. Buyers paying six figures are paying for something more than words: records that show how work got done, linked across systems, over years.
What makes company records worth more
Across buyer pages, the same factors come up again and again:
- Years of history. Troveo's pricing guide puts it simply: five years beats five months (source: Troveo, How much is company data worth). Most programs also set a 3-year floor.
- More systems. Mercor says the more tools you share, the higher your payout (source: Mercor, Monetize your data). Email alone is worth less than email plus CRM plus tickets plus accounting.
- Records that connect. A support ticket tied to the email thread, the job, and the invoice shows cause and effect. That linkage is what AI labs use to train systems that do real work.
- Written procedures. micro1 lists SOPs and knowledge bases first among what it buys (source: micro1, data partnerships).
- Clear rights. Troveo's guide warns that data without provable rights is worth nothing to serious buyers (source: Troveo, How much is company data worth). If you can show you own it and can license it, you are worth more.
- Uniqueness. micro1 reserves its $1M+ tier for highly unique data (source: micro1, data partnerships). Specialized industries and unusual workflows are scarcer than generic office email.
What lowers the price
- Large exclusions. Everything you hold back reduces what a buyer receives. That is often the right call (client files, personnel records and anything under NDA should stay out), but it shows up in the price.
- Short or patchy history. A system adopted last year has little to offer.
- Records that live in people's heads. If the work is not written down, it can't be licensed.
- Exclusivity you didn't price. An exclusive license can pay more up front but blocks other sales. Make sure the price reflects that.
- Rights you can't confirm. Platform terms, customer contracts and NDAs can limit what you are allowed to share.
Law, accounting and healthcare are narrower by design
Client files, privileged material, client books and anything derived from a tax return are never in scope, and patient health information falls under HIPAA. Firms in these fields can still license firm-owned material, but the scope and the price will be smaller. See the guides for law firms and accounting firms.
Why buyer calculators disagree, and what they agree on
Several buyers publish estimators. Every one carries a warning. Handshake AI labels its result not an offer (source: Handshake AI, data partnerships), Mode calls its calculator indicative only (source: Mode, data program), and Pewter calls its example an illustration (source: Pewter).
Where two of them published the same company profile, the ranges line up closely:
| Estimator | Example profile | Example range |
|---|---|---|
| Handshake AI | 100 employees, 10 years, US | $383,083 to $896,969 (source: Handshake AI, data partnerships) |
| Pewter (its licensing partner's model) | 100 people, 10 years | $385,000 to $895,000 (source: Pewter) |
The width of those ranges, more than double from bottom to top, is the honest part. The same company can land at either end depending on scope, exclusions and terms. SimpleClosure's CEO makes a related point about headline offers: check who is making the offer and the buyer chain behind it before you trust the number (source: SimpleClosure, How Real Are Those Million-Dollar Data Offers? (Oct 1, 2026)).
How to get a real number for your records
- Inventory what you have. Systems, years in each, and who owns them.
- Decide what stays out. Client files, personnel records, anything under NDA. Write it down before talking to buyers.
- Ask more than one buyer. Prices and terms vary widely, and some buyers will match a competing written offer (Nyne says it will beat a qualifying one (source: Nyne, compare data licensing offers)).
- Compare written offers, not estimates. Price, scope, exclusivity, payment schedule, deletion and liability, side by side.
That is the work Regulr does for sellers. If you'd like a second opinion first, the complete guide to licensing your company data covers the process end to end, and the buyer directory lists every program we track.
How Regulr is paid, and how we write these pages
Regulr is a sell-side advisor. When a business we introduce signs a license, we may be paid a fee by the buyer or through a program's referral terms. We list every program we can verify whether or not it pays us, order is never set by fees, and every figure is the company's own advertised term with a link to where it says so. Regulr has not closed a data deal yet, so nothing here describes our results.
Frequently asked questions
How much is my company's data worth to AI companies?
There is no public price list. Buyer programs advertise ranges that mostly start around $100,000 for an established company that qualifies: micro1 advertises $100K+, $500K+ and $1M+ tiers, Handshake AI advertises $100K to $4M+, and Mode advertises $100K to $5M. Reported deals for shut-down startups have been smaller, in the tens of thousands of dollars. Your price depends on years of history, how many systems you can include, how well records connect, and what has to stay out.
How much do AI companies pay for data?
Publicly reported prices range from about $0.001 per word for bulk text (Reuters, 2024) to $10 million for bankrupt Spirit Airlines' records, which Google agreed to buy in August 2026. For operating businesses, the programs we track advertise six-figure starting points. Those are advertised figures, not guaranteed offers.
What makes company data more valuable to AI buyers?
Buyers point to the same factors: more years of history, more systems included, records that link together (a ticket tied to the email thread and the invoice), written procedures, clear rights to license what you share, and fewer exclusions. A larger team usually means more records, which is why buyer estimators start from headcount and years.
Are the data value calculators on buyer websites accurate?
They are rough guides, and the buyers say so. Handshake AI labels its estimate not an offer, Mode calls its calculator indicative only, and Pewter calls its example an illustration. Only a written offer that states price, scope, exclusivity and payment schedule tells you what your records are worth to that buyer.
Do I need a minimum number of employees to sell company data?
Most programs do. The common floor is about 20 full-time employees and 3 years of records (Handshake AI, Nyne, Pewter, Mode). micro1 asks for 30+. Mode accepts law firms from 6 and accounting firms from 10. A few programs, such as HUMXN and Troveo, publish no minimum.
Is a startup's data worth less than an established company's?
Usually, yes, because buyers pay for history and depth. Reported shutdown sales through SimpleClosure have mostly been in the tens of thousands of dollars, such as roughly $25,000 for a codebase plus about $15,000 pending for workspace data in one example MIT Technology Review described. Programs aimed at operating companies with years of records advertise higher starting points.
Does this checker give me a price?
No. It matches your business against each program's published eligibility and shows what each program advertises, then scores the factors buyers say drive value. Regulr does not publish its own price estimates, because the honest answer depends on what you can license and on written terms.
Sources
Program terms checked against each company's own pages on October 15, 2026.
- micro1, data partnerships
- micro1, company referral
- Mercor, Monetize your data
- Handshake AI, data partnerships
- Nyne, sell company data
- Nyne, compare data licensing offers
- Mode, data program
- Pewter
- Yahoo Finance, Oct 13, 2026
- Endless
- Troveo, business data
- aitrainer.work summary of Project Lazarus
- Centific, enterprise data
- Polyshares
- Anthroly
- SimpleClosure AssetHub
- MIT Technology Review, Oct 13, 2026
- Sunset blog, Apr 14, 2026
- DataSupply, how it works
- SHFA.ai
- Avelence
- Data Licensing Advisors
- Sell Business Data
- Petrarch
- The Next Web, on the Spirit data sale
- Fortune, Sep 14, 2026
- Forbes, Apr 16, 2026
- Reuters, via KFGO, Apr 2024
- Troveo, How much is company data worth
- SimpleClosure, How Real Are Those Million-Dollar Data Offers? (Oct 1, 2026)
Brian Boesen · Founder, Regulr AI
Brian runs Regulr AI, which works for established businesses that want to license de-identified records to AI companies. He checks each program's published terms against its own pages and updates these guides when they change. Last updated October 15, 2026.
This page is general information, not legal advice. Rules vary by state and by profession, and they apply to your facts, not to a category. Before you share anything, have your own counsel (and for law firms, your state bar's ethics hotline) confirm what is in scope.
Keep reading