Cloudflare now blocks AI training bots on ad pages for new sites

A change Cloudflare announced in July took effect on September 15. New sites on its network now start with AI training bots blocked on the pages that earn money from ads.

What the new default does

From September 15, 2026, domains newly added to Cloudflare block AI training bots and AI agent bots by default on pages that show ads. Search bots are still allowed by default.

There is a catch for the biggest crawlers. Crawlers that combine search with training, such as Googlebot, Applebot and Bingbot, are handled by their most restrictive purpose, so the training block applies to them.

Cloudflare announced the change in a July 1 blog post.

Why ad pages

Cloudflare’s reasoning is that an ad signals a page meant for a human visitor. When an AI system reads that content and answers questions with it, the publisher can lose the visit and the ad impression that pays for the page. Setting the block as the default on those pages means a new site owner does not have to know about AI crawlers to be protected from them.

What to watch

Cloudflare says more than 20% of web domains sit behind it, so a default setting on its network shapes how much new content AI training and agent bots can reach. The distinction it draws, blocking training and agent bots while allowing search bots, also puts pressure on AI companies to keep those crawler categories clearly separated.