# AI training data vendors

URL: https://aidataforsale.com/compare/ai-training-data-vendors
Publisher: AI Data For Sale (Marketing Signals Ltd)
Updated: 2026-11-10

AI training data vendors compared: Scale AI, Surge AI, Mercor, Appen, Labelbox and Prolific, by model, workforce, focus and published figures.

The vendors that create, label and evaluate training data with human workforces, compared on published facts.

## In short

- Vendors differ in model: finished data (Scale AI, Surge AI, Appen), expert time (Mercor), tools plus experts (Labelbox) and participant panels (Prolific).
- Reported revenue in 2024 ranged from about $870 million (Scale AI) to $1.2 billion (Surge AI).
- None of them licenses existing company archives as a core business.

## Compared

| Vendor | Model | Focus | Published figure |
| --- | --- | --- | --- |
| [Scale AI](https://aidataforsale.com/companies/scale-ai) | Finished data | Labeling, expert data, defense | About $870 million revenue (2024) |
| [Surge AI](https://aidataforsale.com/companies/surge-ai) | Finished data | RLHF, RL environments, language | $1.2 billion revenue (2024) |
| [Mercor](https://aidataforsale.com/companies/mercor) | Expert time | Domain experts for labs | $10 billion valuation (2025); 30,000+ experts |
| [Appen](https://aidataforsale.com/companies/appen) | Finished data | Multilingual collection and annotation | $113.4 million revenue (H1 2024) |
| [Labelbox](https://aidataforsale.com/companies/labelbox) | Tools plus experts | Labeling platform and Alignerr | Not published |
| [Prolific](https://aidataforsale.com/companies/prolific) | Participant panels | Human evaluation and research | Not published |

Sources: [SiliconANGLE](https://siliconangle.com/2025/06/08/report-meta-holding-talks-10b-investment-data-labeling-startup-scale-ai/), [Wikipedia](https://en.wikipedia.org/wiki/Surge_AI), [TechCrunch](https://techcrunch.com/2025/10/27/mercor-quintuples-valuation-to-10b-with-350M-series-c/), [Staffing Industry Analysts](https://www.staffingindustry.com/news/global-daily-news/appen-h1-operating-revenue-falls-with-google-contract-termination), [Labelbox](https://labelbox.com/blog/announcing-labelbox-alignerr-connect-with-proven-ai-experts/), [Prolific](https://www.prolific.com/model-evaluation).

## Need data that already exists?

Vendors create data. If you need real-world archives, see [broker vs labeling vendor](https://aidataforsale.com/compare/data-broker-vs-data-labeling-vendor) or [request data](https://aidataforsale.com/request-data).

## Read next

- [Scale AI alternatives](https://aidataforsale.com/compare/scale-ai-alternatives)
- [Expert networks for AI training](https://aidataforsale.com/compare/expert-networks-for-ai-training)

## Questions

### What does an AI training data vendor do?

It creates, labels or evaluates data for AI developers using a human workforce, usually to the buyer's specification.

## Sources

- [SiliconANGLE: Meta reportedly in talks on $10B+ investment in Scale AI (June 2025)](https://siliconangle.com/2025/06/08/report-meta-holding-talks-10b-investment-data-labeling-startup-scale-ai/)
- [Wikipedia: Surge AI](https://en.wikipedia.org/wiki/Surge_AI)
- [TechCrunch: Mercor quintuples valuation to $10B with $350M Series C (October 2025)](https://techcrunch.com/2025/10/27/mercor-quintuples-valuation-to-10b-with-350M-series-c/)
- [Staffing Industry Analysts: Appen H1 operating revenue falls with Google contract termination (August 2024)](https://www.staffingindustry.com/news/global-daily-news/appen-h1-operating-revenue-falls-with-google-contract-termination)
- [Labelbox: Announcing Labelbox Alignerr Connect](https://labelbox.com/blog/announcing-labelbox-alignerr-connect-with-proven-ai-experts/)
- [Prolific: Human evaluation for AI models](https://www.prolific.com/model-evaluation)