Every value below is a company claim or a court or regulator record with its check date, never a confirmed earning, compared as one pairing within AI training data licensing.
No company wins by fees. The one criterion the sources decide is whether a published route for company data exists: Scale has one, Snorkel does not. Declining both is a complete outcome.
Snorkel AI and Scale AI Side by Side, Dated
Each cell quotes the company's own page or a court or regulator record with its check date, as of November 9, 2026 unless a later re-read is dated in the cell. The table carries no score, rank or winner.
| Attribute | Snorkel AI (Snorkel AI profile) | Scale AI (Scale AI profile) | Checked | Source |
|---|---|---|---|---|
| Legal entity, country | Snorkel AI, Inc.; United States (San Francisco, CA address in its privacy policy) | Scale AI, Inc., a Delaware corporation; affiliate Smart Ecosystem, Inc. dba Outlier AI; United States | 2026-11-16 / 2026-11-09 | snorkel.ai/privacy-policy; outlier.ai Terms of Use |
| Self-description | "frontier AI data lab"; "the datasets, environments, evaluation frameworks, and custom solutions"; founded 2019 out of the Stanford AI Lab | human-data and data-labeling company; Scale Data Partnerships, Outlier, RL Environments with 2,000+ tasks | 2026-11-16 / 2026-11-09 | snorkel.ai/company; scale.com |
| Company data program | none on record | Scale Data Partnerships | 2026-11-09 | snorkel.ai; scale.com/data-partnership |
| Stated value to a data holder | none stated | "$10K–$1M+ illustrative value per data partnership, scaling with cadence" | 2026-11-09 | same |
| Payment timing and structure | no program | not published; "opportunity for recurring revenue" | 2026-11-09 | same |
| Eligibility and data wanted | no program | 40+ employees preferred, case by case; warehouses, chat, email and calendar, documents, work tools, business platforms | 2026-11-09 | scale.com/data-partnership |
| Process and anonymization | no program | application or scoping call; anonymized inside the company's environment before extraction; per-package approval; on-prem or VPC | 2026-11-09 | same |
| License terms published | no program | no: exclusivity, term, deletion, retraining not published | 2026-11-09 | same |
| Receiving labs or clients named | none in the sources checked | none on the program page | 2026-11-09 | program pages |
| Referral offer for introducers | none found | up to USD 50,000, only if Scale signs; one referral per deal; repeat intros to the same company not paid | 2026-11-09, re-read 2026-11-16 | snorkel.ai; scale.com/data-partnership |
| Contributor terms | not recorded | via Outlier: hourly or per task, paid weekly, "Rates vary"; work product assignment, individual arbitration | 2026-11-09 | outlier.ai/faq; Terms of Use |
| Own size statement | Series E USD 350 million, valuation USD 3.5 billion, run rate USD 375 million, 2026-09-22 | Meta minority stake, valuation over USD 29 billion, 2025-06-12 | 2026-11-09 | Snorkel blog; Scale blog |
| Court and complaint record | none in the sources checked | McKinney settled (final approval 2026-10-30); Schuster voluntarily dismissed; Ramey notice of settlement; 5 noyb GDPR complaints | as of 2026-11-09 | per record table |
Licensing Company Data: No Program on Record Against a Published Offer
On company data the two records do not line up term by term: Scale states a program, a value range and a referral bonus, and Snorkel states none of them. The sections below take each side in turn.
What Snorkel Lists Instead: Built Datasets and Environments, No Intake for Records
Snorkel's company page names "the datasets, environments, evaluation frameworks, and custom solutions" it delivers as a frontier AI data lab, checked 2026-11-16, and lists no program, form or terms for licensing records a company already holds, checked 2026-11-09. It calls this offering "Data-as-a-Service" and names no client in the sources checked; its software is not reviewed here.
An approach that offers to license a company's records "to Snorkel" therefore cannot be matched to any published route. Its sender, contracting party and terms have to be checked against snorkel.ai, the domain Snorkel uses.
With no Snorkel route on record, such outreach is tested with the red-flag checks that decide whether a data licensing email is legitimate.
What Scale Publishes for Data Holders, and What It Leaves Out
Scale states "$10K–$1M+ illustrative value per data partnership, scaling with cadence", with 40 or more employees as a preferred starting point decided case by case, checked 2026-11-09. It names data in warehouses, chat, email and calendar, documents, work tools and business platforms.
Scale describes a scoping call, anonymization inside the company's environment before extraction and per-package approval gates; the full process sits on the Scale profile. The page leaves out what a seller would sign: payment timing beyond "opportunity for recurring revenue", any clause on exclusivity, term, deletion or retraining, and the receiving labs.
The range is an illustration, not a quote or a payment. Each program's stated values as of 2026-11-09 are set side by side under data partnership payouts and program terms compared.
Referral Bonus for Introducers: None Found Against Up to USD 50,000
Scale states "Refer a company and earn up to $50,000 bonus", paid only if Scale signs a partnership agreement and once per deal, while no referral offer for companies was found for Snorkel, both checked 2026-11-09. A re-read on 2026-11-16 confirmed that "Multiple referrals or intros to the same company will not be paid out", and Scale reviews referrer profiles "within 1–2 business days".
Snorkel's site shows no referral offer, intake route or bonus for introducing a company in the sources checked.
Every other published offer is compared under referral programs of AI data buyers.
Expert Contributors: No Recorded Snorkel Terms Against Outlier's Published Terms
Scale acquires contributor work through Outlier, whose Terms of Use name Scale AI, Inc. and its affiliate Smart Ecosystem, Inc., while the sources checked record no contributor terms for Snorkel as of 2026-11-09. Each value in the table is the platform's own claim.
| Term | Snorkel AI | Outlier, Scale AI (Outlier AI profile) | Checked | Source |
|---|---|---|---|---|
| Contracting entity named | not recorded | Scale AI, Inc. and affiliate Smart Ecosystem, Inc.; Terms of Use effective 2026-05-11 | 2026-11-09 | outlier.ai/legal/terms-of-use |
| Stated pay | not recorded | "Competitive pay + quality-based rewards, paid weekly"; hourly or per task; "Rates vary by expertise, project complexity, and location" | 2026-11-09 | outlier.ai/faq |
| Payment | not recorded | weekly, processed Tuesdays, via PayPal, AirTM or ACH | 2026-11-09 | outlier.ai/faq |
| Screening and eligibility | not recorded | skill screenings, ID and location verification; 117 allowed locations in guidelines effective 2026-07-15 | 2026-11-09 | outlier.ai; flexible-working-guidelines |
| Contract clauses | not recorded | irrevocable assignment of work product (s.4(b)); no withholding except as required by law (s.3); individual AAA arbitration, class waiver, 30-day opt-out for US residents (s.10) | 2026-11-09 | outlier.ai/legal/terms-of-use |
| Paid-out and size claims | none recorded | 900K+ experts; "$500M+" paid to experts (home page) vs "$600M+" paid out (about page) | 2026-11-09 | outlier.ai; outlier.ai/about-us |
Nothing here can be ranked: Outlier publishes no rate, and no Snorkel rate, schedule or contracting arrangement is recorded. Outlier's two paid-out totals disagree, so both stay side by side as its claims.
For a contributor, the clauses weigh more than any total: work product is assigned irrevocably and disputes go to individual arbitration.
Stated rates across platforms and fields are compared under pay for AI training work.
Court and Complaint Record of Snorkel AI and Scale AI
Scale's record holds a court-approved contributor settlement, a voluntarily dismissed contributor suit, a suit with a joint notice of settlement and five GDPR complaints, while no court case, breach notice or regulator action involving Snorkel appears in the sources checked as of 2026-11-09. The table lists each record by date; allegations are not findings.
| Date | Company | Record | Forum / docket | Status as of 2026-11-09 | Source |
|---|---|---|---|---|---|
| n/a | Snorkel AI | none recorded: no court case, breach notice or regulator action involving Snorkel AI, Inc. | n/a | n/a | court and regulator sources checked |
| 2025-01-17 | Scale AI | Schuster v. Scale AI, Inc. et al. filed (contributor class action; allegations) | N.D. Cal. 3:25-cv-00620-JSC | voluntarily dismissed by notice dated 2026-08-10 (with prejudice as to named plaintiffs, without as to the putative class) | Dkt. 34 |
| 2025-02-11 | Scale AI | Ramey v. Scale AI, Inc. et al., joint notice of settlement (Dkt. 32); claims not recorded | N.D. Cal. 3:24-cv-06999 | HireArt stipulation for dismissal 2025-05-12 (Dkt. 36); no dismissal order on the feed | docket feed |
| 2026-10-27 | Scale AI | 5 GDPR complaints by noyb and three partner organizations over Outlier (trackers, Persona biometric checks, access requests) | data protection authorities in France, Italy, the Netherlands and Austria | no authority decision recorded | noyb.eu |
| 2026-10-30 | Scale AI | McKinney v. Scale AI, Inc. et al. (contributor misclassification class settlement), gross USD 12,500,000; payments reported on Form 1099, not W-2 | Superior Court of California, County of San Francisco, CGC-24-620481 | settled; final approval order filed 2026-10-30 | final approval order |
Every Scale record concerns contributors, and none concerns a company data deal. McKinney ended in a court-approved settlement, not a finding on the claims; Schuster was voluntarily dismissed, not settled; the Ramey feed shows no dismissal order.
Snorkel's empty row is a scoped finding, not a clean bill: it covers the sources checked as of 2026-11-09, not matters never filed or published.
Full narratives of the contributor cases sit under contributor lawsuits against human data platforms.
What Each Company States About Its Own Size: Funding Claims, Not a Ranking
Snorkel AI states that it raised a USD 350 million Series E at a USD 3.5 billion valuation and passed an annualized revenue run rate of USD 375 million, in a post dated 2026-09-22, while Scale AI announced on 2025-06-12 that Meta would hold a minority of its outstanding equity at a valuation of over USD 29 billion.
Snorkel names Insight and S32 as the round's leads (confirmed 2026-11-16); a run rate is not audited revenue. Scale states no stake percentage or investment amount.
The figures differ in kind and date, and neither company's revenue or headcount is stated in the sources checked, so neither is called the larger.
Sector figures are gathered under funding and valuations of AI data companies.
Which Data Holder or Contributor Each Company Fits
Neither company is better by fees; each one's own pages point to different readers, and only Scale addresses a company holding data, as of 2026-11-09.
Scale's stated terms point to:
- Companies of 40 or more employees with data in warehouses, chat, email, work tools and business platforms.
- Holders who want anonymization in their own environment and approval per package.
- Introducers who accept a bonus paid only if Scale signs.
Snorkel's record points to:
- No data holder: Snorkel publishes no route for company records.
- Contributors only after they check Snorkel's own current terms, since none are recorded here.
Neither may suit:
- A holder that needs published license terms or named receiving labs first.
- A holder that must secure deletion or retraining rights in writing from the start.
Buyers beyond these two are weighed under choosing a buyer for your data.
Licensing to Scale AI, or Neither: Benefits, Risks, Downsides and Alternatives
For a data holder the choice here is Scale or neither, since Snorkel publishes no route for company records, and licensing to Scale means signing terms that are not public in advance. The lists draw on the record as of 2026-11-09.
Benefits
- A published process with anonymization inside the company's environment.
- Per-package approval gates before any data leaves.
- A stated value range and eligibility, both dated claims.
Risks
- No published license scope, exclusivity, term, deletion or retraining terms.
- Contributor lawsuits and GDPR complaints on Scale's record, none about data partnerships.
- Outreach invoking Snorkel has no program to match.
Downsides
- The range is illustrative, not a quote.
- No payment timing or structure is published.
- Snorkel offers no route for company data at all.
Alternatives
- Other published programs, among them Handshake AI Data Partnerships, Labelbox Prism, Mercor, micro1 Company Data Partnerships and Turing Project Lazarus.
- An intermediary or sell-side adviser.
- An enquiry for an independent introduction.
- Not licensing the data at all.
Keeping the data is a complete outcome, and every option is laid out under alternatives to selling data to AI labs.
Where This Comparison Sits Among the Buyer and Human Data Pages
This page sets two human-data companies against each other; each profile holds the full record, and the directory places them among every other buyer. The parent directory answers who buys AI training data.
- Every program: company data partnership programs.
- Snorkel's other pairings: Snorkel AI vs Mercor and Snorkel AI vs Handshake AI.
- Scale's other pairings: Scale AI vs Surge AI and Turing vs Scale AI.
- Competitors: Scale AI alternatives and dataset vendors and data labs.
- Case records: McKinney v. Scale AI and noyb complaints against Scale AI.
- Contributor side: human data for AI and RL environment companies.
How This Comparison Is Checked and Kept Independent
Each cell is read from the company's own page, a court filing or a regulator record and dated; a cell changes only when its source has been read again.
Snorkel comes first only by URL and H1 order, and inclusion, order and tone depend on neither company, as the disclosure sets out. Court summaries are information, not legal advice.
Frequently Asked Questions
Does Snorkel AI buy company data like Scale AI?
No program for it is on record: Snorkel lists none, while Scale runs Scale Data Partnerships, both checked 2026-11-09.
How much does Snorkel AI pay contributors compared with Outlier?
Neither can be compared: no Snorkel rate is recorded in the sources checked, and Outlier states "Rates vary by expertise, project complexity, and location", checked 2026-11-09.
Applications and assessments are compared under AI training jobs and platforms.
Which is bigger, Snorkel AI or Scale AI?
No primary source compares them: Snorkel states a USD 375 million run rate (2026-09-22) and Scale a valuation of over USD 29 billion (2025-06-12), figures of different kinds and dates.
Can a company license data to both Snorkel AI and Scale AI?
Only Scale has a published route; Scale publishes no exclusivity terms, so the signed agreement would decide what else the company may license.
What does each company do?
Snorkel describes itself as a frontier AI data lab delivering datasets, environments and evaluation frameworks; Scale is a human-data company that also licenses company data and runs Outlier.