Cloudflare Says 50% of Internet Traffic Is Non-Human as 50+ AI Content Deals Emerge
Updated
Updated · The Cloudflare Blog · Jul 1
Cloudflare Says 50% of Internet Traffic Is Non-Human as 50+ AI Content Deals Emerge
1 articles · Updated · The Cloudflare Blog · Jul 1
Summary
More than 50% of internet traffic is now non-human, Cloudflare said, arguing that a monetized market for web content has emerged as publishers lose referral traffic to AI-driven discovery.
52% of crawler requests were for AI training in June 2026, up from 22% in spring 2025, while mixed-use crawlers made up more than 36%, blurring whether content is being indexed for search or reused for AI.
More than 50 publisher-AI licensing agreements have been signed since 2023, which Cloudflare said followed its default blocking of AI training crawlers on new domains unless site owners opt in.
Google remains a key friction point because it drives about 88% of referral traffic yet uses a mixed-purpose crawler, making it harder for publishers to separate search visibility from AI access and compensation.
Cloudflare, which says more than 20% of the web runs behind its network, plans more attribution, freshness and market-discovery tools as it pushes for bot self-identification and scalable content licensing.
Will AI licensing deals save the open web, or only create a new class of media giants?
If creating content no longer brings traffic, will the quality of information powering future AI models inevitably decline?
Over 50% of Web Traffic Is Now AI Crawlers: How Publishers Are Fighting Back, Monetizing Access, and Reshaping the Open Web
Overview
Since June 2026, the web has entered a new era where automated agents and AI bots generate most online traffic. This machine-majority web means AI is now the main way people access information and shop online, fundamentally changing how content is discovered and used. While this shift brings new opportunities, it also creates big challenges for publishers. Many businesses that rely on ads or subscriptions are threatened as AI systems use their content without compensation, leading to an uneven relationship where content is widely consumed but publishers see little return. As a result, traditional business models are at risk.