Deal explainer
Tumblr, WordPress.com and the Automattic AI data deals
Quick answer
In February 2024, reporting revealed that Automattic, the parent company of Tumblr and WordPress.com, was nearing deals to supply user post data to OpenAI and Midjourney for AI training. Internal documents showed the company had already compiled a large, and in places overbroad, batch of content before adding a public opt-out setting.
What was reported#
- 404 Media reported internal documents showing Automattic compiled an 'initial data dump' of all public Tumblr posts from 2014 to 2023 for OpenAI and Midjourney.
- The compiled dataset reportedly mistakenly included private posts, deleted blogs, unanswered asks and other content that was not supposed to be shared.
- Automattic published a statement titled 'Protecting User Choice' and rolled out a toggle letting users opt out of third-party AI data sharing.
- The Verge reported the deals with OpenAI and Midjourney were described internally as 'imminent' at the time of the reporting.
- Neither OpenAI nor Midjourney publicly confirmed financial terms of any resulting agreement.
Related deals
SourceX was not a party to these deals. This tracker summarizes public reporting so businesses can see how AI data licensing is developing.
Sources
Public sources, checked 2026-10-09. SourceX is not affiliated with the organizations named.
- 404 Media: Tumblr and WordPress to Sell Users' Data to Train AI Tools 404media.co
- The Verge: Tumblr's owner is striking deals with OpenAI and Midjourney for training data theverge.com
- Engadget: Tumblr and WordPress posts will reportedly be used for OpenAI and Midjourney training engadget.com
See if your company qualifies
A short company assessment. No data uploads are needed.