Cloudflare now blocks AI crawlers by default and launches pay-per-scrape program

Cloudflare has switched to blocking AI crawlers by default for its customers and launched a Pay Per Crawl program that lets website owners charge AI companies for scraping access. The move represents a significant shift from the previous free-for-all approach to AI data collection, potentially forcing major AI companies to negotiate and pay for content access rather than scraping without permission.

What you should know: Over 1 million Cloudflare customer websites had already activated the company’s AI-bot-blocking tools before this default change took effect.

• Cloudflare can identify even “shadow” scrapers that aren’t publicly disclosed by AI companies, using behavioral analysis, fingerprinting, and machine learning to distinguish between AI bots and legitimate crawlers.
• The company’s customers can still opt out of the blocking if they want to allow unrestricted AI scraping of their content.

Why this matters: AI scraping has intensified to levels that can mimic DDoS attacks, straining servers and potentially knocking websites offline.

• Many publishers, particularly news organizations, want AI companies to pay for using their content rather than scraping it for free.
• The shift could “dramatically change the power dynamic” between AI companies and content creators, according to Nicholas Thompson, CEO of The Atlantic and former WIRED editor-in-chief.

The bigger problem: Traditional web protection methods aren’t working against aggressive AI scrapers.

• The widely used Robots Exclusion Protocol (robots.txt) is frequently ignored by AI companies, with over 26 million scrapes violating the protocol in March 2025 alone, according to Tollbit, a content licensing platform.
• “Robots.txt is ignored,” says Danielle Coffey, president and CEO of the News Media Alliance, which represents several thousand North American outlets.

What they’re saying: Industry leaders see this as a necessary step toward fair compensation for content creators.

• “We firmly believe that all content creators and publishers should be compensated when their content is used in AI answers,” says Bill Gross, CEO of AI startup ProRata, which has agreed to participate in the Pay Per Crawl program.
• “We’ve been feverishly trying to protect ourselves,” Coffey explains about news organizations’ efforts to combat unauthorized scraping.

The reality check: Major AI players haven’t yet committed to the Pay Per Crawl program, which remains in beta.

• Companies like OpenAI have struck licensing deals with publishers, including WIRED’s parent company Condé Nast, but specific terms haven’t been disclosed.
• An entire ecosystem of tutorials exists online teaching web scrapers how to evade Cloudflare’s blocking tools, suggesting the arms race will continue.