New: who buys company data, October 2026 list
Data Licensing Report
Get offers

Buyers directory

Who buys company data for AI training? (Updated October 2026)

As of October 2026, 25 providers publicly invite businesses to license or sell their internal records for AI. Each is listed alphabetically on the same fields.

Get offersHave an offer? Compare it

By Updated Why trust usHow we make money

We may earn a referral fee when a business we introduce completes a deal. Telegraph Lab is a commercially affiliated provider.

Data Licensing Report may earn a referral fee when a business that applies through this site completes a deal with a participating provider. Telegraph Lab is a commercially affiliated provider: the owner of Data Licensing Report is paid commission on some Telegraph Lab deals, and does not own Telegraph Lab. Providers are listed alphabetically and described from their own public materials using the same fields.

How we make money

How many people work at your company?

A provider is listed when it says publicly that it will pay a business for internal records such as email, chat, documents, business-software records, recordings or code, for use in AI, either itself or through the buyer it introduces. The type label under each name shows how each one works. General data marketplaces and programs for media or medical data are listed separately under other channels.

Every row is re-checked monthly, and sooner when news changes a provider's program. Before going further with any provider, compare its terms with at least one other: send your company profile to our team for review, and the offer benchmark lays out an offer you already have.

ProviderWhat they buyWho they buy fromHow they payPublished rangesLast checkedNext step
Appen[1]Licenses directlyProcess documents, decision and approval records, systems of recordOperating and winding-down companiesFixed paymentPriced per deal; paid when the data is acceptedGet offersProfile
Avelence[2]IntroducerChat, email, documents, CRM, support and other system records, through its data partnersOperating companies; No published minimum to submit a profileThe matched data partner pays you; Avelence may be paid by that partnerCalculator output, e.g. $160K–$500K at default inputsGet offersProfile
Corpus[3]Licenses directlyEmail, chat, meeting transcripts, documents, tickets, CRM notesOperating companies; 25+ or 30+ employees (its pages differ)Fixed paymentOne-time payment on delivery confirmation; no revenue share$100K–$2M+ by headcountGet offersProfile
DataVendor (HUD)[4]Listing marketplaceCodebases, data bundles and task setsMainly shutting-down startups' codebasesSeller sets the price and receives 80% of the payoutExample listings $68K–$220K; $5K baseline per codebaseGet offersProfile
Defined.ai[5]Licensing agentAudio, video, image, text and multimodal datasetsBusinesses that generate or aggregate dataRevenue shareLicensing fees from marketplace salesAverage partner revenue $600K+ (Defined.ai's figure)Get offersProfile
FileYield[6]Listing marketplaceDataset listings in 36 categories, including chat logs and helpdesk ticketsOperating companiesBuyer pays the seller directly; FileYield charges sellers 8%Published guide and catalog prices; not seller payoutsGet offersProfile
Google Content Offer Pilot[7]AI company, directCode repositories and other specialized, non-public contentDevelopers and small businesses; Google invites as space allowsFixed paymentFee agreed with Google, paid for delivered contentGet offersProfile
Handshake AI[8]Licenses directlyEmail, files, chat and business-software records20+ full-time employees and 3+ years operatingTerms scoped per deal under NDA$100K–$4M ($4M+ on its program page)Get offersProfile
idler[9]Licenses directlyCodebases and de-identified business dataOperating companies; 10–100 employeesLicenses your data; no payment terms publishedGet offersProfile
LH2 AI Labs[10]Licenses directlyCode, workflow records, recorded calls and meetingsEnterprises, hospitals and factoriesNo payment terms for data owners publishedGet offersProfile
License My Data[11]IntroducerOperational documents and workplace records, placed through processorsOperating companies; 30–200+ employees; or 100+ with 4+ years across 3+ systems in useFixed paymentPer-buyer non-exclusive licenses or a one-time exclusive buyout“Six and seven figures” (its description)Get offersProfile
Mercor[12]Licenses directlySlack, Teams, email, documents and other tool data, through connectorsOperating companies; 8 listed industries, from legal to logisticsFixed paymentWire on delivery, typically 2–4 weeks after signing; no feesGet offersProfile
micro1[13]Licenses directlyCRM data, project histories and process documentationOperating companies; 30+ employees (its application page says 30–200)Paid according to the value of the data$100K–$2M+Get offersProfile
Miro Advisory[14]IntroducerOperating datasets and private codebasesOperating companies; About 20–700 employees with 5+ years of historyAdvisor only; how it and sellers are paid is not published$100K–$1M+ (indicative)Get offersProfile
Nyne[15]Licenses directlyAccounting, CRM, support, chat, email and document recordsOperating companies; 20+ U.S. W-2 employees and 3+ years of workflowsFixed paymentPaid once the dataset is delivered and acceptedTypically $100K+Get offersProfile
Polyshares[16]Licenses directlyMessaging, email, files, CRM, tickets, design files and callsAny company with real operating historyFixed paymentPaid directly, net 30–60 after approval; no seller fee$100K–$2M+Get offersProfile
Replay[17]Licenses directlyCode, internal documents, communications, financial and operating recordsOperating companies; Any industry; price tiers start at 20 employeesFixed paymentRevenue shareTypically upfront cash plus a perpetual revenue share$10K–$1M+ by headcountGet offersProfile
Scale AI[18]Licenses directlyData warehouse, chat, email and business-system recordsOperating companies; Prefers 40+ employeesPartnerships designed for recurring revenue$10K–$1M+ per partnership (illustrative)Get offersProfile
Scrimdata[19]Licenses directlyInboxes, Slack, tickets, documents, CRMs and calendarsOperating, pivoting, winding-down or exited companiesPrice agreed in each contract; no amounts publishedGet offersProfile
Sell My Business Data[20]IntroducerEmail, documents, orders, support and workflow recordsOperating businesses in listed industriesBuyers pay you; its own fee at closing is unstated in amount and payerGet offersProfile
SimpleClosure (AssetHub)[21]Wind-down platformSource code and workspace data (Slack, email, Drive, Notion)Closing companies; operating companies since September 2026Fixed paymentSells or licenses assets; SimpleClosure, an affiliate or an introduced buyer may buyTypically $10K–$100K per closing company (Forbes, April 2026)Get offersProfile
Sunset[22]Wind-down platformSlack threads, product specs and code repositoriesStartups winding downFixed paymentBuys datasets outrightSimilar to SimpleClosure (Forbes)Get offersProfile
Telegraph Lab[23]Licenses directlyAffiliatedWork conversations, business-software records and drive filesOperating companies; No minimum stated for direct applicants; referral pages cite 20+ or more than 30 employees and 3+ yearsFixed paymentPrice and payment milestones set in a written agreement$100K–$4MGet offersProfile
Troveo[24]Licensing agentChat, email, documents and business-system records; also video and audioOperating and closing companies; No published minimum size or ageRevenue shareRevenue share on every sale; no fees“Six figures” typical for full-company deals (its description)Get offersProfile
Turing (Project Lazarus)[25]Licenses directlyCodebases, design docs, test logs and wikis from startupsDefunct startups, per its launch announcementHUD, which runs a competing codebase marketplace, wrote in June 2026 that 'Current public pricing shows $10K–$100K for legacy code and up to $15K per modern repository.'Get offersProfile

Scroll the table sideways to see every column.

Each row is drawn from the provider's own public pages, using the same fields for every provider; the number beside each name links to that page in the sources below. Names link to our profile where one exists. "Get offers" opens the application, which comes to our team for review. A dash means no entry is recorded yet for that field, not that the provider withholds it. The methodology sets out the inclusion test, update schedule and corrections process.

Which providers have been compared side by side?

Other channels you may see mentioned

  • AWS Data Exchange

    Amazon's catalog for data files, tables and APIs sold to AWS customers.[26]

  • Databricks Marketplace

    Marketplace for datasets, models and apps shared with Databricks users.[27]

  • Datarade

    General B2B data marketplace where providers list data products for any buyer.[28]

  • FirstParty

    Data-commercialization firm for large enterprises; productizes and distributes data rather than licensing internal records for AI.[29]

  • Protege

    Licenses health, video, audio and other datasets for a revenue share on each use; no program aimed at companies' internal records was found.[30]

  • Search.co Data Marketplace

    Listing marketplace for business, web and AI-training datasets.[31]

  • Segmed

    Revenue-share licensing of de-identified medical imaging from healthcare providers.[32]

  • Snowflake Marketplace

    Marketplace for data products sold to Snowflake customers; providers need a Snowflake account.[33]

  • Versos AI

    Helps video libraries license video datasets to AI labs.[34]

Sources

  1. Appen program page, Appen. Accessed .
  2. Avelence program page, Avelence. Accessed .
  3. Corpus program page, Corpus. Accessed .
  4. DataVendor (HUD) program page, DataVendor (HUD). Accessed .
  5. Defined.ai program page, Defined.ai. Accessed .
  6. FileYield program page, FileYield. Accessed .
  7. Google Content Offer Pilot program page, Google Content Offer Pilot. Accessed .
  8. Handshake AI program page, Handshake. Accessed .
  9. idler program page, idler. Accessed .
  10. LH2 AI Labs program page, LH2 AI Labs. Accessed .
  11. License My Data program page, License My Data. Accessed .
  12. Mercor program page, Mercor. Accessed .
  13. Monetize your company's operational data (micro1), micro1. Accessed .
  14. Miro Advisory program page, Miro Advisory. Accessed .
  15. Nyne program page, Nyne. Accessed .
  16. Polyshares program page, Polyshares. Accessed .
  17. Replay program page, Replay. Accessed .
  18. Scale AI program page, Scale AI. Accessed .
  19. Scrimdata program page, Scrimdata. Accessed .
  20. Sell My Business Data program page, Sell My Business Data. Accessed .
  21. SimpleClosure (AssetHub) program page, SimpleClosure (AssetHub). Accessed .
  22. What are RL gyms? How founders can monetize startup data in a wind-down (Sunset blog), Sunset. Published . Accessed .
  23. License Your Business Data | Telegraph Lab, Telegraph Lab. Accessed .
  24. Troveo program page, Troveo. Accessed .
  25. Turing (Project Lazarus) program page, Turing (Project Lazarus). Accessed .
  26. AWS Data Exchange, AWS Data Exchange. Accessed .
  27. Databricks Marketplace, Databricks Marketplace. Accessed .
  28. Datarade, Datarade. Accessed .
  29. FirstParty, FirstParty. Accessed .
  30. Protege, Protege. Accessed .
  31. Search.co Data Marketplace, Search.co Data Marketplace. Accessed .
  32. Segmed, Segmed. Accessed .
  33. Snowflake Marketplace, Snowflake Marketplace. Accessed .
  34. Versos AI, Versos AI. Accessed .

Get offers

Find the next step for your company’s data.

  • No records or exports needed to apply
  • Free for businesses
  • Your company profile comes to our team for review

Already have an offer? Compare it

Start your application

Do you keep at least 3 years of email or business-software history?

We may earn a referral fee if a deal closes. How we make money