A restaurant changes its menu prices. An AI search crawler could keep downloading the page to check, or receive a signal that something changed. Cloudflare’s chief executive says OpenAI is paying for that second approach.

Based on 4 reviewed sources · See sources

Paying for the signal

On Cloudflare’s 5 November earnings call, co-founder and CEO Matthew Prince described hints about changed pages as a way to reduce repeated crawling. He said OpenAI pays Cloudflare for that service. The company had announced the research partnership in July; this is a payment statement about that work, rather than a new search launch.

“OpenAI is paying for that”

Matthew Prince · Cloudflare CEO, 5 November earnings call · Read the original ↗

Prince gave no price, payment start date or publisher revenue share in that answer. His broader prediction that small payments could become a major business remains a forecast.

Being found and being chosen

The companies’ 15 October pilot report separates finding a page from ranking it. Cloudflare supplies network signals; OpenAI independently evaluates and ranks pages. The research uses web metadata for search and discovery, rather than sharing page content for model training. A signal also does not grant permission to access a page.

The report says the next phase will measure what happens after discovery. A crawler finding a shop’s updated menu is therefore not, by itself, evidence of a citation, a customer visit or a sale.

A different payment for access

Cloudflare’s separate Pay Per Crawl service lets site owners set a price for AI crawlers to access content. Its current documentation still labels it a closed beta. That is a different proposition from selling freshness signals to OpenAI.

For publishers and businesses, the unresolved question is what reaches them: less unnecessary server work, more useful visitors, compensation, or some combination. The reviewed material does not quantify those outcomes or disclose a publisher payout from the signal arrangement.

Sources

Written with AI assistance from the sources above. Analysis and practical implications are our interpretation. Our editorial approach.