Updated November 2026
Scale AI Data Partnerships: what the seller program offers
Scale AI's Data Partnerships program pays companies to license operating records with names and personal details stripped out, and gives $10K-$1M+ as illustrative value per partnership. Scale prefers 40+ employees. Its page publishes no license terms, payment timing or named sellers, and Meta holds 49% of Scale.1,2,3,4,5
Its page says identities never leave, and also says chat keeps participants "preserved intact" and files keep authorship.6
Its referral bonus appears as "up to $50,000" in the FAQ and "$10,000" on the referral tab of the same page.7
Key facts
Scale AI in its own words and others'
| Fact | Value and source |
|---|---|
| Scale AI: deal size | $10K-$1M+, labeled "illustrative value per data partnership, scaling with cadence" |
| Scale AI: who it takes | "40+ employees" as its preferred starting point, assessed case by case, plus years of records in existing tools, real workflows, domain depth and a clear internal owner |
| Scale AI: handling | Anonymizes inside your environment before extraction and strips PII; nothing is shared without per-package approval; encryption in transit and at rest, a full audit trail, and on-prem and VPC options |
| Scale AI referral program | FAQ: "Refer a company and earn up to $50,000 bonus". Referral tab: "Refer a company and earn $10,000", paid when Scale signs a partnership agreement |
| Ownership | Meta holds 49% after a deal reported at $14.3B; founder Alexandr Wang joined Meta |
| Scale AI leadership | Francis deSouza became CEO effective 2026-08-10; Jason Droege was interim CEO from June 2025; Alexandr Wang is quoted as Founder and Chairman of the Board |
How much does it pay?
Scale AI says $10K-$1M+, and calls that illustrative
What it buys
What Scale AI wants and who it takes
Scale AI wants records from warehouses such as BigQuery and Snowflake, Slack, Teams and Google Chat, email and calendar, documents, Jira, Confluence, Asana, Notion and Linear, and platforms such as Salesforce, HubSpot, Workday, ServiceNow, SAP and Zendesk. Its named industries run from software and law firms to freight brokerage, property management, staffing and hospitality.1,9
Scale calls "40+ employees" its preferred starting point and assesses each case. It also wants years of records in existing tools, real workflows, domain depth and a clear internal owner. No year count or region is stated.3
Do names leave?
Scale AI's page conflicts on whether identities leave
Scale AI says employee and customer identities never leave, and the same page says chat keeps participants "preserved intact", email keeps senders and recipients, and files keep authorship. The page does not say whether those fields are tokenized. Ask for a sample record before anything else.6
Scale says it anonymizes inside your environment before extraction and shares nothing without per-package approval, with encryption, a full audit trail, and on-prem and VPC options.10
What is missing?
What Scale AI's program page does not say
Scale AI's page does not state whether a deal is a license or a sale, and gives no exclusivity terms, retention period, payment timing or seller fees. Those are the terms that set what a deal is worth, so get them in writing before you scope any data.4
Who gets the data?
Meta's 49% stake and Scale AI's buyers
Scale AI is 49% held by Meta after a deal reported at $14.3B, and founder Alexandr Wang joined Meta, while Scale's blog still quotes him as Founder and Chairman of the Board. Interim CEO Jason Droege wrote that Scale remains independent, and Francis deSouza became CEO effective August 10, 2026.5,11,12
Scale's FAQ says AI companies will not use your data to compete because they are modeling personas. It names no buyers and does not mention Meta. In our view, ask directly whether Meta can receive your data.8
Referrals
Scale AI's referral program shows two figures
Scale AI's FAQ says "up to $50,000" for referring a company, and its referral tab says "$10,000", paid when Scale signs a partnership agreement. Both figures are live on one page, and we found no written referral terms.7
Process
How it works with each buyer
In each buyer's own words. Where a buyer publishes nothing, we say so.
Scale AI
- Structured, de-identified records from warehouses (BigQuery, Snowflake, Databricks), chat (Slack, Teams, Google Chat), email and calendar, documents, work tools (Jira, Confluence, Asana, Notion, Linear) and platforms (Salesforce, HubSpot, Workday, ServiceNow, SAP, Zendesk)1
- "40+ employees" as its preferred starting point, assessed case by case, plus years of records in existing tools, real workflows, domain depth and a clear internal owner3
- $10K-$1M+, labeled "illustrative value per data partnership, scaling with cadence"2
- Anonymizes inside your environment before extraction and strips PII; nothing is shared without per-package approval; encryption in transit and at rest, a full audit trail, and on-prem and VPC options10
- Not published
Check first
How to check Scale AI, or any buyer
The same checks for every company we profile. We state facts; you decide.
- 1.Find the company behind the ad: a legal name, a registered address and named people.
- 2.Ask for a written offer and a sample agreement before you send any data.
- 3.Ask who removes personal data, by what method, and approve it before anything is exported.
- 4.Ask which AI labs will receive the data, and whether you will be told.
- 5.Ask how they are paid: by you, by the buyer, or from a share of your deal.
- 6.Get terms from at least one other buyer before you answer.
Reader reviews
What people who worked with Scale AI say
7.0/10
None yet
A written summary appears once enough reviews are published. It will quote both praise and complaints.
No reviews yet. Be the first.
If you have sold data to, hired, or worked for Scale AI, your review helps the next person decide.
FAQ
Questions, answered with sources
Scale AI calls 40+ employees its preferred starting point and assesses each partnership case by case. No year count or region is stated.3
Scale AI's page says identities never leave, but also says participants, senders, recipients and authorship are kept. Ask how those fields are handled and see a sample.6
Scale AI's page shows two figures: "up to $50,000" in its FAQ and "$10,000" on its referral tab. Both pay only once Scale signs.7
Every business is different, so this one is best talked through with the buyer you're working with. Their team goes over it with you before anything is signed.
Before you sign
Questions to talk through with your buyer
Every business is different, so these are best talked through with the buyer you're working with. Their team goes over them with you before anything is signed.
Questions for this kind of business
- Is a Scale AI data partnership a license or a sale, and can it be exclusive?
- Can Meta, or any company Meta controls, receive our data?
- Are participant, sender and author fields tokenized before they leave our environment?
- How do CCPA, GDPR, UK GDPR and state privacy laws apply, and is de-identified data still personal information?
- Do customer, supplier or platform contracts, NDAs or API terms restrict this?
Questions for any buyer
- Is the license exclusive, and for how long?
- Can you sublicense the data, and to whom?
- Which AI labs will receive it, and will you tell us?
- How long do you keep it, and how is it deleted?
- What is your liability cap, and who indemnifies whom?
- Will the contract ban re-identification of people in the data?
- Who removes personal data, by what method, and do we approve it first?