Who's buying data
Companies and AI-data vendors that publicly pay organizations for their data: what each pays, the data it wants and who qualifies.
Amounts come from each program's own published terms as of Oct 8, 2026. They are ranges, ceilings and rates, so a listed amount is not an offer. Every program here is run by the company that pays for the data, or by an agent that signs the license with the seller. Consumer-data buyers and affiliate middlemen are left out. Who Sells Data isn't paid for any listing and doesn't use referral links: every source links to the program's own public page.
Download as JSON · 38 programs from 29 companies
- Pays companies for data
- Licenses or buys data from an organization and pays it directly.
- Referral bounty
- Pays whoever introduces a company that goes on to sign.
- Licensing agent or marketplace
- Licenses an owner’s data to other buyers, or runs a marketplace, and pays the owner a share.
- Invite-only
- No public terms; the company invites partners or screens interest, and decides whom it works with.
38 of 38 programs
| Company | Program | What they pay | Data they want | Who qualifies | Source |
|---|---|---|---|---|---|
| Appen | Enterprise Data Pays companies for data | Quoted per partnership; no amounts publishedAppen covers extraction, anonymisation and transfer costs and removes personal data. It also sells enterprise company data to AI labs. | Operational documentation, decision and escalation records, and systems of record. | Companies. Those winding down get a faster path. | appen.comobserved Oct 8, 2026Quote from AppenTerms are agreed in writing before extraction begins, and payment is on acceptance |
| Brickroad | Supplier program Licensing agent or marketplace | Split set per license, in writing; no supplier feeBrickroad acts as a licensing agent rather than a buyer. Its buyers lean toward alternative financial data. No split percentage is published. | Data packages licensed to AI labs, funds, corporates and agents. | Data suppliers; onboarding is in beta. | brickroad.networkobserved Oct 8, 2026Quote from BrickroadWe do not buy your data, and we are not owned by anyone who does. |
| Centific | Enterprise Data Pays companies for data | No amounts published; paid on acceptanceCentific covers extraction, anonymization and transfer costs. Its healthcare area includes claims and de-identified clinical documentation. | Operational data in finance and accounting, healthcare operations, sales and deal workflows, people operations, and IT service management. | Enterprises, including companies winding down. | centific.comobserved Oct 8, 2026Quote from CentificWe want to partner with enterprises to anonymize and license your operational data for AI training by frontier labs. And we’ll pay you for it. |
| Collinear AI | Data acquisition Pays companies for data | No amounts published; an initial range is quoted within a dayA one-time export under a license; ownership stays with the company. Collinear uses the data to build training and evaluation environments. | Code repositories and business app history, such as pull requests with the tickets behind them. Personal data is removed. | Companies. Prices rise with 15+ people at peak, several years of history, human-written code, and work visible across several tools. | collinear.aiobserved Oct 8, 2026Quote from Collinear AILicense your code repositories and business app history for AI training and evaluation. |
| Defined.ai | Partnership Programs: Data Partner Licensing agent or marketplace | Licensing fees from marketplace sales; no rate publishedThe $600K+ average partner revenue on the page is Defined.ai's own figure. There is no fixed minimum dataset size. | Audio, video, image, text or multimodal data. Cinematic, anime, coding and STEM data are in high demand. | Any business that generates or aggregates data, with consent and the right to commercialize it for AI training. | defined.aiobserved Oct 8, 2026Quote from Defined.aiMonetize your data ethically and easily with Defined.ai’s Partnership Programs. |
| Partnerships to improve our AI products Invite-only | Paid pilot for non-public content; no public rates or applicationThe page is undated. Press reports of other Google data purchases are not on this page and are not listed here. | Closed and offline datasets, enhanced metadata and signals, and real-time structured facts. Science, mathematics and learning are priorities. | No public application route. | ai.googleobserved Oct 8, 2026Quote from GoogleFor example, we’ve been piloting a partnership program to pay for the delivery of non-public content in a range of media formats. | |
| Halluminate | Data Partnerships Invite-only | Interest form; no terms publishedThe page does not say whether Halluminate buys, licenses or shares revenue. | Company data and workflows. Types are not specified. | Halluminate reviews contact details and decides whether there is a fit. | halluminate.aiobserved Oct 8, 2026Quote from HalluminateTurn your company’s data and workflows into a new revenue opportunity. |
| Handshake AI | Data Partnerships Pays companies for data | Published range $100K to $4M+Handshake's license page (joinhandshake.com/ai/data-partnerships/license/) gives the range as $100K to $4M and states the 20+ full-time employees and 3+ years rule. Payment typically follows 30 to 60 days after Handshake accepts the de-identified dataset. | Company data such as business process flows, hiring logic and document-level communications, scoped and de-identified. | Companies with 20+ full-time employees and 3+ years in operation. | joinhandshake.comobserved Oct 8, 2026Quote from Handshake AIWe partner with you to license your company data. Partnerships can range from $100K to $4M+ depending on the data shared. |
| Handshake AI | Refer a company Referral bounty | $10,000 per referred company that signsNo separate referral terms are published, so rules on who can refer and on clawbacks are not stated. | Introductions to companies that could license their data to Handshake AI. | Referred companies need 20+ full-time employees and 3+ years in operation. | joinhandshake.comobserved Oct 8, 2026Quote from Handshake AIYour $10K bonus is paid when their deal closes |
| HUD (DataVendor) | DataVendor marketplace Licensing agent or marketplace | Sale proceeds, less a 20% fee on bank withdrawalsIn beta, and fees are current settings that can change. Withdrawals need $1,000 before fees, and earnings settle 14 days after payment. | Codebases, raw data such as documents, logs and enterprise records, tasksets, and RL environments sold with a HUD taskset. | Organizations can sell; individuals cannot. Vendor setup is self-serve with no application review, and buyers are invited. | docs.datavendor.aiobserved Oct 8, 2026Quote from HUD (DataVendor)DataVendor's 20% platform fee applies only when a vendor withdraws to a bank. |
| Labelbox | Prism data licensing Pays companies for data | From about $50K to $1M+ per licenseLicenses are exclusive by default. Prism's individual categories, such as consented likeness, voice and medical records, are outside this listing. | Real-world operational data from companies. | Companies. Prism also sources from individual contributors, which this entry does not cover. | prism.labelbox.comobserved Oct 8, 2026Quote from LabelboxPrism acquires real-world operational data from companies and contributors to power frontier AI while you retain full control of your data and rights. |
| Magma | Data licensing for AI product companies Licensing agent or marketplace | Revenue share as the data licenses; no percentage publishedMagma licenses the data to labs and researchers. Its site does not describe anonymization; handling and retention are left to the data agreement. | Interaction data from AI agents: traces, corrections and completed workflows. | Companies building AI agents. | magmahq.aiobserved Oct 8, 2026Quote from MagmaA revenue share, paid as your data licenses. Terms are in the data agreement we sign during onboarding. |
| Mercor | Mercor Data Pays companies for data | No amounts published; pay depends on data volume and the tools connectedPaid by wire when the anonymized dataset is delivered to the buyer, typically 2 to 4 weeks after signing. Anonymization is automated, with no manual review. The page's healthcare examples are clinical notes, claims and intake forms. | Data from business tools such as Slack, Teams, Gmail, Outlook, Salesforce, GitHub, Confluence, Drive and Notion (100+ supported). | Businesses in legal, healthcare, finance and insurance, professional services, manufacturing, retail, real estate, and logistics. | mercor.comobserved Oct 8, 2026Quote from MercorThe data your business generates every day is exactly what frontier labs need. We license it, package it, and pay you for it. |
| Mercor | Refer a company Referral bounty | Referral bonus; amount not publishedThe public page names no amount. An amount appears inside Mercor's signed-in referral app, so it is not listed here. The bonus is paid when the referred company goes live and starts earning. | Introductions to businesses that could become paid Mercor Data partners. | Not stated on the public page. | mercor.comobserved Oct 8, 2026Quote from MercorRefer them to Mercor. When they become a paid partner, you get rewarded too. |
| micro1 | Enterprise Data Partnership Pays companies for data | Published tiers of $100K+, $500K+ and $1M+The tiers are floors. micro1's company-referral page lists company compensation of $100K to $2M+ for approved data packages. Personal data is removed when relevant. | Operational data and documentation from established companies. Higher tiers cover data across several teams and business functions. | Operationally mature companies with 30+ employees. US companies are prioritized, then other Western markets. | micro1.aiobserved Oct 8, 2026Quote from micro1Compensation is based on the quality, uniqueness, and value of the data provided |
| micro1 | Refer a company Referral bounty | Up to $50,000 per referred companyRaised from $25,000 in a June 2026 micro1 announcement. Payouts are at micro1's discretion, can be clawed back, and are credited to a micro1 account wallet. | Introductions to companies that could license their operational data to micro1. | Anyone over 18; the referral terms exclude micro1 employees and most contractors. Referred companies: 30+ employees, US prioritized, primarily English. | micro1.aiobserved Oct 8, 2026Quote from micro1You can earn up to $50K per referral, depending on the size and value of the partnership. |
| Microsoft | Publisher Content Marketplace Licensing agent or marketplace | Paid on delivered value; no rates publishedAnnounced February 3, 2026; no newer status was found. The post covers grounding answers and does not describe model training. | Premium publisher content, licensed for grounding answers in AI products. | Publishers, through an interest form. The program is a pilot. | about.ads.microsoft.comobserved Oct 8, 2026Quote from MicrosoftThe result is a direct value exchange: publishers will be paid on delivered value, and AI builders gain scalable access to licensed premium content that improves their products. |
| Mode (data.mode.inc) | Company data partnerships Pays companies for data | Indicative range $100K to $5MThis is Mode at data.mode.inc, not Mode Analytics. Mode's calculator labels the range indicative only. Mode's main business pays consumers for personal data; only its company program is listed here. About three months from first conversation to payment, with no broker fee. | An agreed copy of company-owned operating records: email, documents, chat, CRM and billing data. | Strongest fit is 20+ full-time U.S. office employees; accounting firms 10+, law firms 6+. Nonprofits are eligible. | data.mode.incobserved Oct 8, 2026Quote from Mode (data.mode.inc)Your company’s data could be worth $100K–$5M. |
| Mode (data.mode.inc) | Referral program Referral bounty | $50,000+ per qualified company (6% of the deal or $50,000, whichever is greater)Terms from Mode's Affiliate Referral Agreement v1.0 (published Oct 6, 2026): 6% or $50,000 for the first five qualified deals, rising after 25; companies headquartered outside the US earn 6% with no minimum. Paid within 30 days of Mode's payment, with a 12-month clawback. | Introductions to businesses with valuable operating data. | Counts only when the referred company signs a full corpus license within 6 months. | data.mode.incobserved Oct 8, 2026Quote from Mode (data.mode.inc)Know a business with valuable operating data? Make the introduction. Earn when the partnership qualifies. |
| Nyne AI | Data partnerships Pays companies for data | Six-figure payoutsNyne's core business is an identity graph of people; only its company data program is listed here. Paid after acceptance. | Company workflow data as it exists today: chat, email and calendar, documents, and CRM, HR, ERP and support records. Nyne removes identifying details on its side. | 20+ U.S. W-2 employees and 3+ years of workflows. | nyne.aiobserved Oct 11, 2026Quote from Nyne AIBefore your data is used, identifying details are removed and records are cleaned and formatted. This work happens on our side, you transfer the data as it exists today, and we take it from there. |
| Nyne AI | Refer a company Referral bounty | Up to $20,000 per introduced company that completes a transactionNo written referral terms are published; referrals are made by email. | Introductions to companies with 20+ U.S. full-time employees. | The referred company must have at least 20 U.S. full-time employees and complete a transaction with Nyne. | nyne.aiobserved Oct 8, 2026Quote from Nyne AIEarn up to $20,000 for each company you introduce that has at least 20 U.S. full-time employees and completes a transaction with Nyne. |
| Ooak Data | Partner with us Pays companies for data | Up to $300K; closed deals range from $20K to several hundred thousand dollarsThe ceiling and deal range are Ooak's own figures. Raw data is deleted within 30 days of delivery under the contract. | Company data from work tools. Value depends on years of history, headcount, and how many tools connect and how deeply. | Operating, winding-down and closed companies. Payment goes to the company entity. | ooakdata.comobserved Oct 8, 2026Quote from Ooak DataDeals to date range from $20K, the smallest closed, to several hundred thousand dollars. |
| Opendatabay | Licensed data marketplace Licensing agent or marketplace | Sellers keep 70% to 95% of each sale, rising with transaction sizeListing is free. Commission also applies to off-platform deals with introduced buyers for up to 24 months, and sellers cannot grant an exclusive licence over delivered files for 12 months. The fee schedule was last updated October 6, 2026. | Text, image, audio, video, code, agentic trajectories, 3D spatial, tabular, time-series, human feedback and synthetic data. | Registered businesses and organizations, after identity checks. Individuals cannot sell. | opendatabay.comobserved Oct 8, 2026Quote from OpendatabayListing on Opendatabay is free. Opendatabay applies a tiered commission structure on Self-Serve Sales, API calls, and data access fees. |
| Polyshares | Data licensing Pays companies for data | $100K to $2M+ per licenseAll figures and case studies are Polyshares' own; no press or funding coverage was found. Paid Net 30 to 60 once the data is approved, anonymized and delivered; no fee. | Company data licensed directly from the company that produced it, in any sector. | Companies in any sector. Offers within 1 to 2 weeks. | polyshares.comobserved Oct 8, 2026Quote from PolysharesWe buy it. Polyshares is the counterparty on the deal. |
| pre.dev | Code Licensing Licensing agent or marketplace | Standing rates of $100 per million source tokens plus $40 per million git-history tokensStanding rates are not a minimum. Payment comes within 15 days of pre.dev receiving the buyer's funds, and buyers commonly take 60 days to pay. The agreement also lets pre.dev take its own paid licence at the standing rates. | Private codebases with their git history. Public repositories are not eligible. | Owners of private repositories. AI labs choose which repositories to license. | pre.devobserved Oct 8, 2026Quote from pre.devAt current standing rates: $100 per million source tokens plus $40 per million git-history tokens, for one standard non-exclusive buyer licence. |
| Protege | For Data Providers (outside healthcare) Licensing agent or marketplace | Revenue share on qualified deals; no percentage publishedProtege's healthcare program pools de-identified patient records and is not listed here. | Data and content outside healthcare. The provider page does not list types. | Organizations that own their data. Partners keep ownership. | withprotege.aiobserved Oct 8, 2026Quote from ProtegeWe formalize licensing through a clear revenue share agreement and structure sustainable terms with AI buyers. |
| Scale AI | Data Partnerships Pays companies for data | Illustrative $10K to $1M+ per partnershipScale labels the range illustrative; it scales with delivery cadence. Partnerships can recur. | Company data, anonymized inside the company's own environment before extraction, with approval for each package. | 40+ employees preferred; Scale assesses each partnership case by case. | scale.comobserved Oct 8, 2026Quote from Scale AIYour data becomes an asset you're paid for |
| Scale AI | Refer a team Referral bounty | Up to $50,000 per referred companyNothing is paid unless Scale signs a partnership agreement. No separate referral terms are published. | Introductions to companies that could become Scale data partners. | Referrals go through Scale's referral form. One payout per deal, to whoever made the successful introduction. | scale.comobserved Oct 8, 2026Quote from Scale AIRefer a company and earn up to $50,000 bonus. |
| Shutterstock | Data Partner Network Licensing agent or marketplace | No payout terms publishedContent is used for training and evaluation and is not resold in Shutterstock's marketplace. | Content libraries, delivered as a full library, a curated collection or part of custom training datasets. | Content owners, through a lead form. | shutterstock.comobserved Oct 8, 2026Quote from ShutterstockNo. Participation in Shutterstock’s Data Partner Network does not involve transferring ownership of your content. |
| SimpleClosure | AssetHub Licensing agent or marketplace | Success fee on sales or licenses; past deals ran from tens to hundreds of thousands of dollarsSimpleClosure brokers sales and does not publish its percentage. Its September 3, 2026 press release says AssetHub is free for SimpleClosure dissolution customers. The page does not mention consent from employees whose messages are in workspace data. | Startup source code and workspace data. Sensitive employee and personal information is excluded. | Startups, including companies still operating. | simpleclosure.comobserved Oct 8, 2026Quote from SimpleClosureSimpleClosure only takes a percentage if your assets are successfully sold or licensed. |
| SuperAnnotate | Data Partnerships Licensing agent or marketplace | Indicative valuation $100K to $1M+Paid on delivery and on each relicense. Rights are licensed for a set period to buyers the company approves. | Business records built up over time that the company has the rights to license. | Companies of any size, industry or location. | superannotate.comobserved Oct 8, 2026Quote from SuperAnnotateSuperAnnotate prepares and licenses your operational data to frontier AI labs, so you earn from it without the work. |
| The General Data Company | Data Partnerships Pays companies for data | Purchase price or license fees agreed per deal; no amounts publishedPayment follows milestones set in a written agreement. The page's estimator is labelled "not an offer". | Operational data from European companies, sold outright or licensed. Recurring workflow contributions are paid as agreed fees. | Data owners, authorized representatives, advisers and insolvency administrators. | thegeneraldata.comobserved Oct 8, 2026Quote from The General Data CompanyWe buy and license operational data from European companies. |
| The General Data Company | Referral program Referral bounty | Up to $50,000 per introductionThe reward depends on the size of the partnership and is set in a separate written referral agreement. | Introductions to companies whose data The General Data Company can license. | Anyone who registers for a personal referral link. Paid when the introduction leads to a signed data partnership. | thegeneraldata.comobserved Oct 8, 2026Quote from The General Data CompanyKnow a company with valuable data? Earn up to $50,000. |
| Troveo AI | Business Data Licensing agent or marketplace | Revenue share on every sale; full-company deals typically start at six figuresTroveo licenses on the owner's behalf, charges no fees and earns from the revenue share. Payout totals and example figures on the page are Troveo's own and are left out here. | Company data exports. | Companies that own their data. No other requirements are stated. | troveo.aiobserved Oct 8, 2026Quote from Troveo AIAI labs and startups are actively paying six figures for company data exports. |
| Troveo AI | Troveo for Content Owners Licensing agent or marketplace | Royalty on every use; no rate publishedOwners keep full ownership. Content categories include talking-person and interview footage. | Video libraries. Featured owners also supply audio, text and gaming content. | Content owners, from media companies to creators. | troveo.aiobserved Oct 8, 2026Quote from Troveo AIEvery use earns you a royalty. |
| Turing | Project Lazarus Pays companies for data | A single payment of up to $1MTuring confirms the figure after sampling, usually within a week. The seller keeps ownership; Turing acquires defined rights to the prepared record. | A company's operating record: communications, tickets, wikis, knowledge bases, databases and source code. Code is not bought on its own. | Operating companies, and teams winding down or already closed. | lazarus.turing.comobserved Oct 8, 2026Quote from TuringFrontier AI labs train on how real companies operate, not the polished product. Lazarus turns that operating record into a single payout of up to $1M, and handles everything else. |
| Turing | Project Lazarus referral Referral bounty | Up to $10,000 per referred companyThe cap appears on the homepage button "Refer & Earn up to $10k". No referral terms page exists. | Introductions to companies with years of operating data. | No eligibility rules are published. | lazarus.turing.comobserved Oct 8, 2026Quote from TuringKnow a company sitting on years of operating data? Send them our way and earn when it becomes a deal. |
| Versos AI | Video training data marketplace Licensing agent or marketplace | No rates, fees or revenue split publishedVersos runs the software and marketplace; AI labs and model developers are the buyers. | Video libraries, prepared into datasets for AI labs and model developers. | Video library owners, through a demo or contact form. Content stays in the owner's storage. | versos.aiobserved Oct 8, 2026Quote from Versos AITrusted by the world’s leading video libraries to prepare and deliver bespoke datasets to frontier AI labs. |
The list is unranked; sorting only reorders. Sorting by an amount puts programs without a stated upper limit last. Each amount is the program's own figure on the date shown; follow the source to check current terms.