Automated internet traffic has surpassed human traffic for the first time.

With the booming development of AI Agents this year and the continuous influx of various new AI startups, automated crawlers and intelligent agents have completely flooded the internet. According to the latest Radar telemetry from Cloudflare, the world's largest content delivery network (CDN) provider, bot traffic has officially surpassed human traffic for the first time in history, signaling that the future of internet traffic will likely be heavily dominated by AI.

Long before the birth of modern generative AI, a substantial portion of internet traffic was already automated, typically consisting of search engine indexers, malicious scripts, vulnerability scanners, and brute-force attacks. However, with the ongoing generative AI boom, an increasing number of AI companies are deploying massive fleets of web scrapers to crawl the global internet to harvest information for model training. Because training a cutting-edge Large Language Model (LLM) requires astronomical amounts of web data, AI-driven automated traffic has seen a dramatic spike over the past couple of years.

Furthermore, moving into this year, the rapid rise of Agentic AI has enabled many users to self-host proxy applications like Openclaw. Features widely integrated across various AI platforms—such as real-time web search and the Model Context Protocol (MCP)—have further accelerated this automated traffic. Today, advanced agents can completely replace humans in operating web browsers to perform the vast majority of tasks, including bypassing CAPTCHAs, logging into accounts, and automatically managing social media profiles.

Cloudflare notes that these AI agents are primarily busy reading product descriptions, tracking real-time prices, and executing complex, multi-step online tasks. These include comparing flight options, scraping and indexing web content (specifically tailored for AI models rather than traditional search engines), acting as personal assistants to order food, comparing retail deals, and even independently handling backend customer service workflows.