AI has stolen publishers’ revenue

Tech’s ubiquitous adoption of generative AI clients has crippled publishers.

The Big Picture: Bots crawl publishers’ web pages and summarize their content without referring users to those pages. As a result, publishers receive no traffic or ad revenue from this engagement, forcing them to rethink their business models in order to survive.

Between the Lines: Cloudflare CEO Matthew Prince outlined the problem with some astounding figures during a recent event at Cannes.

• A decade ago, Google used bots to crawl web pages but only did so at a ratio of 2:1.
• In January 2025, following the widespread adoption of AI, Google raised that proportion to 6:1. OpenAI and Anthropic were far less generous to publishers, crawling 250 and 6,000 web pages, respectively, for every referral to a publisher.
• This month, the trend has accelerated wildly, with Google’s ratio spiking to 18:1, OpenAI’s to 1,500:1, and Anthropic’s to a stunning 60,000:1.

Conclusion: For publishers, ad revenue often depends on click-through rates, or the percentage of users viewing a webpage who click on an ad. Without visits to their web pages, publishers get no clicks and, therefore, no money. As people trust AI bots and their summaries more, they’re less likely to seek out publishers’ web pages on their own.

Prediction: Cloudflare has already announced that they’re working on tools that would prevent bots from scraping websites. Look out for that tech, as well as any potential lawsuits that tech firms might bring against them as they develop AI browsers.