# Directory of 187 companies that buy and sell AI training data

> Original title: We launched a directory of RL/data companies that buy and sell AI training data http://whosellsdata.com

- Source: X search: AI startup funding (we raised, seed, Series A)
- Published: 2026-11-04T01:25:51.000Z
- Bad Influence: https://badinfluence.news/items/tuumn451el7al121ow2z0jv1s
- Original: https://x.com/ben_cerchio/status/2117790396420124974

## Summary

A directory tracking companies that buy and sell AI training data now lists 187 firms, up from 159 at launch. This week, Turing's Project Lazarus changed from a single payment of up to $1M to a revenue share on every sale. The list adds Protege AI, which has raised $65M, and sfdata, which raised a $30M seed round led by Khosla.

## Post

We launched a directory of RL/data companies that buy and sell AI training data http://whosellsdata.com

187 companies now, up from 159 at launch (biggest update yet), every fact sourced.

This week: Turing's Project Lazarus moved from a single payment of up to $1M to a revenue share on every sale. That means the seller's data keeps paying as long as it keeps selling

New on the list:
- @withprotegeai (@BobbySamuels @engyziedan): $65M raised, licensed healthcare, video and audio data
- @sfdata (@brandonguo @faizsiddiqii):  $30M seed led by Khosla, AI production traces turned into training data
- @GLG Frontier (@gradygardner): credentialed experts who test AI models
- @CintGroup (@comerpatrick): human text, audio and video responses
- @wattdata (@howdymaudi): $4.3M pre-seed, licenses private company data
- @XploreAi (@Carlos_Bort): RL envs and evals for physical engineering
- @aiastrolabe (@babarkkhan @haidarkk1): human-made data and red teaming for Arabic dialects
- @ConsequenceLabs: custom human-generated datasets
- Miro Advisory (@Robi_Feldman): helps firms license records and code
- Telegraph Lab (@KennyBerg_): licenses business records to AI labs
- Mode (@novaes_ai): buys company records and resells them as training data
- ReadySolve: MartechBench, an RL env for marketing work
- Navaday: expert-made data and licensed business records

Who's missing?